{ "$schema": "folio-bean-index/v1", "tile": { "beans": { "count": 132, "unit": "open beans" } }, "repoWeb": "https://github.com/litlfred/folio-assistant", "fileWeb": "https://github.com/litlfred/folio-assistant", "fileRef": "cat/cat-harness/beans", "items": [ { "id": "folio-assistant-015u", "title": "The fixed '▾ Folio' handle covers page titles and content", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T10:36:13Z", "updatedAt": "2026-09-24T17:45:58Z", "preview": "The Folio glass handle is fixed at the top and overlaps each page's heading, and at 390 px the glass's own first controls. Reserve space for it, or move it out of the reading flow.\n\nObserved on: `external-schemas`, `folio`, `library`, `navbar`, `tools` (see each `cat-harness/docs/wireframes//intent.md`). Per-page detail is in each visualiser's task under the epic.\n\n## Done when\n\n- [x] No pag", "file": "beans/defs/folio-assistant-015u--the-fixed-folio-handle-covers-page-titles-and-cont.md" }, { "id": "folio-assistant-03nl", "title": "MERGE-MAIN NOTIFICATION HYGIENE: one bad matrix member reds the whole run, and an unchanged failure re-notifies on every push to main", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T05:52:09Z", "updatedAt": "2026-10-07T17:38:00Z", "preview": "## Brief\n\n**What and why.** The owner is being emailed *\"Merge main into opted-in PRs /\nmerge:main into PR (1958) — Failed in 6 minutes and 18 seconds\"* many times a\nday. `merge-main.yml` fires on every push to `main` (many per hour) and runs one\nmatrix member per opted-in PR; a single member's failure makes the whole run\n`failure`, and GitHub emails the run. Run `37179860536` failed on exactly on", "file": "beans/defs/folio-assistant-03nl--merge-main-notification-hygiene-one-bad-matrix-mem.md" }, { "id": "folio-assistant-065p", "title": "SWEEP TRAP: a stale local `main` ref makes every branch look like it carries unmerged beans", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-25T15:48:01Z", "updatedAt": "2026-09-27T08:27:57Z", "preview": "Found 2026-09-25 during a `/goal-review` sweep, by axis 4's own instruction —\n*\"Items that exist only on a branch … Sweep the open branches' stores too, or\nthe review is blind to the largest workstream.\"*\n\n## The trap\n\nThe natural way to write that comparison is the way the skill's prose implies:\n\n```sh\ngit diff --name-only --diff-filter=A main..origin/ -- beans/defs/\n```\n\nIn this containe", "file": "beans/defs/folio-assistant-065p--sweep-trap-a-stale-local-main-ref-makes-every-bran.md" }, { "id": "folio-assistant-06e3", "title": "docs-auto: a handler at cat-harness/docs-auto// that derives documentation for a sub-graph — and the authoring rule that the author must summarise what it indexes", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-0lmb", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T20:54:07Z", "updatedAt": "2026-10-07T17:15:00Z", "preview": "Owner, 2026-09-20 (session_014HGPQoUnzXGqSspA8x6YyD), in three messages.\n\n## 1. The handler\n\n> in cat-harness needs to be harness/handler at `cat-harness/docs-auto//`\n> defined. which will auto-generate extracatable documentation at `` sub-graph.\n> extracablle = bpmn, tasks, glossary, etc. ther is a glosarry bean... this could\n> clarify it lives at `cat-harness/docs-auto", "file": "beans/defs/folio-assistant-06e3--docs-auto-a-handler-at-cat-harnessdocs-autoauto-do.md" }, { "id": "folio-assistant-073f", "title": "KG-JSONLD STAGING: each stages its JSON-LD on cat//kg-jsonld; CI gates it; the gated branch publishes to the CDN (bootstrap-first walk, failed subgraph skips its cone)", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-whlc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T21:42:43Z", "updatedAt": "2026-10-07T17:37:00Z", "preview": "Owner, 2026-10-02, verbatim: \"i think we also need cat-kg-jsonld in which all of the json(ld) is staged before being published. right now. it is a bit adhoc. bootstrap-tools writes to bootstrap's gh-pages direcrtly i thinkj..... instead (whoever is doing the rendering of the KG):\n* the writer puts 's new/incremental updates to the json(ld) KG in the 'a cat-kg-json-ld branch\n* the CI ch", "file": "beans/defs/folio-assistant-073f--kg-jsonld-staging-each-stub-stages-its-json-ld-on.md" }, { "id": "folio-assistant-08u4", "title": "SURVIVAL PATH: every referenced IRIS node resolves by Handle, with a liveness check and a snapshot pointer", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-kupb", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T18:01:44Z", "updatedAt": "2026-09-30T08:26:13Z", "preview": "From the `v048` roast, objections 6 + 7. Placed under `kupb` by the owner, 2026-09-24.\n\n**Measured:** every catalogue node id is a DSpace UUID or a path, and every upstream URL is `iris.who.int/…`. `hdl.handle.net` occurs **0** times in `who-iris/catalogue`, against `iris-dspace.md` R1: *\"Resolve by **Handle** where one exists… only the Handle is guaranteed outside WHO\"*. No step in `sample-import", "file": "beans/defs/folio-assistant-08u4--survival-path-every-referenced-iris-node-resolves.md" }, { "id": "folio-assistant-0b8c", "title": "Derived artefacts with a branch-kept input are built at publish, never committed (fsh-guts viewer first)", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-nama", "blocking": [], "blockedBy": [], "createdAt": "2026-10-05T18:49:32Z", "updatedAt": "2026-10-06T06:14:10Z", "preview": "Owner, 2026-10-05: *\"how to fix process so wont fail? update process/skills\"*, then option 1 (build at publish), and *\"use dependencies between KG derivations and such (caches, auto-docs, etc)\"*.\n\n## Failure\nThe committed page `cat-harness/docs/fsh-guts/index.md` is derived from `fsh-guts/`, which is stored on `cat/cat-harness/fsh-guts` (keyedBy tip) and moves without a main commit. bean rva2's st", "file": "beans/defs/folio-assistant-0b8c--derived-artefacts-with-a-branch-kept-input-are-bui.md" }, { "id": "folio-assistant-0dav", "title": "QA READERS F2b: the eleven self-sidecar gates compute and judge; audit-coverage stops reporting a moved kind as empty", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T08:47:13Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "Arc `3fva`, from reader audit `gxvk` (`cat-harness/docs/proposals/qa-readers-audit-2026-10-01.md`, family F2b). Refines `oqe3` 3.1c and part of 3.1a. As of 2026-10-06 this is not blocked on `16ei`: it completed (qa-store landed; CI publishes `main/` and `pr/` entries and the prune runs on schedule), so the block is withdrawn. Its write-in-check fixes to `skill-register.ts` and `check-harness-state", "file": "beans/defs/folio-assistant-0dav--qa-readers-f2b-the-eleven-self-sidecar-gates-compu.md" }, { "id": "folio-assistant-0fua", "title": "Long flat lists with no search or filter", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T10:36:14Z", "updatedAt": "2026-09-30T16:48:12Z", "preview": "Pages of 46 to 240+ rows are one flat list with no search, filter or grouping, so finding an entry means scrolling tens of thousands of pixels on a phone. Add a filter and grouping in the shared list component.\n\nObserved on: `glossary`, `processes`, `skills-index`, `tools` (see each `cat-harness/docs/wireframes//intent.md`). Per-page detail is in each visualiser's task under the epic.\n\n## Re", "file": "beans/defs/folio-assistant-0fua--long-flat-lists-with-no-search-or-filter.md" }, { "id": "folio-assistant-0g7s", "title": "processes visualiser: 7 wireframe findings", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T10:36:15Z", "updatedAt": "2026-10-07T17:32:00Z", "preview": "Findings from the as-is wireframe `cat-harness/docs/wireframes/processes/` (intent.md, as-is.html, checks/), observed at 1280×800 and 390×844. Verbatim from its `## Findings`; a finding tagged → is also covered by that cross-cutting bug.\n\n1. **\"Searchable\" in the H1, but the page has no search or filter control.** The only search is the site-wide theme search. Finding one of 68 processes means scr", "file": "beans/defs/folio-assistant-0g7s--processes-visualiser-7-wireframe-findings.md" }, { "id": "folio-assistant-0hi8", "title": "DEPLOY: a local HTTP server that serves a rendering with its declared media types", "status": "completed", "type": "feature", "priority": "high", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T08:55:36Z", "updatedAt": "2026-10-09T14:40:00Z", "preview": "From [#363](https://github.com/litlfred/folio-assistant/issues/363): \"no github,\nlocal git only. tools (several options) to run lightweight server locally for\nskill of hosting content (using mime time data whch github tool cant)\" and\n\"github on private repo so gh-pages not avaialbe. use local server skill.\"\n\n## This hole is already documented, deliberately\n\n`skills/ui/ui-core/serving-renderings.md", "file": "beans/defs/folio-assistant-0hi8--deploy-a-local-http-server-that-serves-a-rendering.md" }, { "id": "folio-assistant-0ipy", "title": "AGENTIC SE LITERATURE: five papers on agentic code review, adversarial debate and agent routing — ingest, analyse within and across, and attack folio-assistant with them", "status": "completed", "type": "epic", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T23:15:14Z", "updatedAt": "2026-10-12T18:26:29Z", "preview": "placeholder", "file": "beans/defs/folio-assistant-0ipy--agentic-se-literature-five-papers-on-agentic-code.md" }, { "id": "folio-assistant-0js9", "title": "INGEST + METHODOLOGY: five subject-indexing papers (SemEval-2025 Task 5 x3, LCSH agentic pipeline, LCSHBench)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-slw1", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T19:46:16Z", "updatedAt": "2026-09-26T12:13:49Z", "preview": "Ingest five open-access subject-indexing papers into `cat-harness/library/`, and adopt the methodology they support. The platform library holds only sources a methodology cites (`library-ref.test.ts`), so the ingest and the methodology nodes have to land together.\n\n## The five sources\n\n| bib-slug | paper | role |\n|---|---|---|\n| `arxiv-2605.03537v1` | Chow, skill-based agentic pipeline for LCSH in", "file": "beans/defs/folio-assistant-0js9--ingest-methodology-five-subject-indexing-papers-se.md" }, { "id": "folio-assistant-0jtl", "title": "SKILLS: a review package in folio-assistant-core — large-document-review, review-heatmap, review-navigation, learned from WHO SOPs and inspection practice", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-q4jm", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T21:02:55Z", "updatedAt": "2026-10-10T16:40:53Z", "preview": "Owner: *\"review ... skills for document review for large things like a DAK or\nL1 digital transofrmation hadbook\"* and *\"put into folio-assistant-core\"*.\n\n**What.** A `review` skill package at `folio-assistant-core/skills/review/`,\ndeclared in folio-assistant-core.json:\n- `large-document-review`: the method. Slice the document by the folio/\n graph, not by page. Assign slices to reviewer roles by s", "file": "beans/defs/folio-assistant-0jtl--skills-a-review-package-in-folio-assistant-core-la.md" }, { "id": "folio-assistant-0kbt", "title": "gen-docs-pages publishes a QA count that includes the witness files it deletes in the same run", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T09:39:59Z", "updatedAt": "2026-10-07T11:50:39Z", "preview": "The `qa` tile and `assets/qa/index.json` are projected BEFORE the orphan sweep removes witness files whose section no longer exists, so one run after a section is renamed or removed publishes a count that includes a file that run itself deleted.\n\nMEASURED on the #1898 branch, 2026-10-04, while bean `vqlp` was being fixed. A section `the-basic-flow` was added, generated (writing `witnesses/document", "file": "beans/defs/folio-assistant-0kbt--gen-docs-pages-publishes-a-qa-count-that-includes-.md" }, { "id": "folio-assistant-0lde", "title": "L1 EVIDENCE: ingest the external evidence an L1 guideline computes from (PICO question sets, Cochrane/WHO systematic reviews, GRADE profiles)", "status": "todo", "type": "task", "priority": "", "parent": "folio-assistant-qvxh", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T10:13:24Z", "updatedAt": "2026-10-04T10:13:24Z", "preview": "Owner, 2026-10-01 (qvxh): smart-L1 is 'fully computable from smart-base assets (and maybe some other things like PICO, cochrane, etc)'. Owner, 2026-10-04 (qvxh option 1): name what is not yet a graph rather than fake one.\n\nFour l1-guideline sections (key-questions, evidence-retrieval-and-synthesis, certainty-of-evidence, evidence-to-decision) are computed from evidence smart-base has not ingested.", "file": "beans/defs/folio-assistant-0lde--l1-evidence-ingest-the-external-evidence-an-l1-gui.md" }, { "id": "folio-assistant-0lj4", "title": "UNDECLARED SIBLING: cat-harness imports 20x from bootstrap-tools, which it does not declare needing — and four checks all report clean", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T11:15:58Z", "updatedAt": "2026-10-08T08:05:00Z", "preview": "`cat-harness/` imports **20 times** from `bootstrap-tools/`, an instance it does\n**not declare needing** — and **four separate checks all report clean.**\n\nFound 2026-09-30 while a sibling agent was planning the `adapters/` closure. It\nflagged the relation as *undetermined* and said it could not find what treats\nthe two as one layer; this bean is that question answered.\n\n## The declarations\n\n```\nca", "file": "beans/defs/folio-assistant-0lj4--undeclared-sibling-cat-harness-imports-20x-from-bo.md" }, { "id": "folio-assistant-0lmb", "title": "CONTENT MODEL: block kinds, adapters and the authoring surface", "status": "in-progress", "type": "epic", "priority": "normal", "parent": "folio-assistant-npuo", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T11:43:44Z", "updatedAt": "2026-10-04T15:12:16Z", "preview": "The block kinds and adapters an author actually writes against.\n\n`55ao` is the one with the largest blast radius — a real `recommendation` kind\nmeans a builder, a Zod schema, a label prefix, viewer registration, constraint\nrows and QA criteria, about thirty files — and it is tracked rather than\nhalf-done for exactly that reason. `cz17` brings the DAK type in properly.\n\n`6xaz` and `bqrg` are the au", "file": "beans/defs/folio-assistant-0lmb--content-model-block-kinds-adapters-and-the-authori.md" }, { "id": "folio-assistant-0mf0", "title": "S2 merge treadmill: land merge:main (#1754) and regen-in-CI on the merge result (d33q part B)", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-7x5n", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T08:14:33Z", "updatedAt": "2026-10-06T19:30:00Z", "preview": "G4. Main moves ~every 3 min; a merge->regen->CI cycle is ~22 min; three handovers stalled on it.\n\n## Done when\n- [x] #1754 merged — 2026-10-01T17:24:57Z\n- [x] d33q part B: CI regenerates on the merge result — merge-main.yml (#1820, #1834), green on main\n- [x] median PR-green -> merged measured before and after — measured 2026-10-06: green→merged median 2.2→5.7 min (not faster); merged-while-red", "file": "beans/defs/folio-assistant-0mf0--s2-merge-treadmill-land-mergemain-1754-and-regen-i.md" }, { "id": "folio-assistant-0mpw", "title": "Remote mount of a harness: declared directory with a remote source, defaults in the harness's own declaration", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-fnx4", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T17:55:54Z", "updatedAt": "2026-10-23T14:13:00Z", "preview": "Owner ruling 2026-10-06 (this session): replace submodule linking (e.g. litlfred/smart-ra has folio-assistant as a submodule) with a REMOTE MOUNT. Not .deps/ (dot-dir; collides with GitHub conventions and our own dot-prefix guard).\n\n## Design (option 1, amended by the owner)\n- A remote mount is a DECLARED DIRECTORY with a new SubgraphSource member, e.g. source: {remote: {repository, ref: <40-char ", "file": "beans/defs/folio-assistant-0mpw--remote-mount-of-a-harness-declared-directory-with.md" }, { "id": "folio-assistant-0qjq", "title": "MERGE-MAIN APPROVAL STALL — CORRECTED: already handled by design; the only gap is `stage`, deliberately preview-only, and the real fix is #1829 D1", "status": "scrapped", "type": "bug", "priority": "low", "parent": "folio-assistant-hfag", "blocking": [], "blockedBy": [ "folio-assistant-9f05" ], "createdAt": "2026-10-02T14:10:55Z", "updatedAt": "2026-10-16T06:30:00Z", "preview": "## This bean was filed on a false premise — corrected 2026-10-02\n\n**I filed it without reading `merge-main.yml` first.** That is the error worth\nrecording, because everything the first version \"discovered\" is written in that\nfile, measured on the same day, with its reason and its fix.\n\nThe step is **\"Have CI judge the merge commit\"**:\n\n> Every GATING workflow, so each judges the merge commit: a bo", "file": "beans/defs/folio-assistant-0qjq--merge-main-approval-stall-the-bots-push-creates-ac.md" }, { "id": "folio-assistant-0r7u", "title": "CONTENT SPLIT 2026-10-06: authoring, review and publication to folio-assistant-core; methods stay in cat-harness — before seeding and before GOAL 5 resumes", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-7x5n", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T06:42:34Z", "updatedAt": "2026-10-07T17:38:00Z", "preview": "Owner, 2026-10-06, verbatim: *\"do content split propoerly across repos\"*, *\"that needs to be done before F\"*, and *\"not a race to separate. do it properly.\"* Placement rule: *\"folio-asst-core focused on content authroiung, review pubilcation. cat-harness on SDLC, tooling, decision making, general methdologies, basic common infra/state mgmt\"*.\n\nStory `rfuq` (S4, direction and placement) owns th", "file": "beans/defs/folio-assistant-0r7u--content-split-2026-10-06-authoring-review-and-publ.md" }, { "id": "folio-assistant-0s6w", "title": "MERGE VERIFICATION: reading merge-tree's rc through a command substitution reports the WRONG exit status, and it reads as clean", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-hfag", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T02:07:04Z", "updatedAt": "2026-10-04T06:03:36Z", "preview": "A merge steward verifies mergeability with `git merge-tree --write-tree` and reads the exit\ncode: **rc=0 clean, rc=1 CONFLICT, rc>=2 an ERROR that is not clean**. On 2026-10-03 that\nread was wrong for a whole session, in one specific shell form, and it reported **clean**.\n\n## The defect\n\n```sh\ngit merge-tree --write-tree origin/main \"$sha\" >/dev/null 2>&1\necho \"merge-tree rc=$?\" ", "file": "beans/defs/folio-assistant-0s6w--merge-verification-reading-merge-trees-rc-through.md" }, { "id": "folio-assistant-0tg5", "title": "STATE BRANCH P4: bean gates run on push to 'state'; harness-dirs, audit:coverage, kg:audit read the storage field", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T10:58:10Z", "updatedAt": "2026-10-09T07:45:00Z", "preview": "Nine bean gates in code-quality-gates.yml move to a state-branch workflow. A gate that cannot fetch the branch reports unknown, never a pass.\n\n## Done when\n- [x] state-branch workflow with the bean gates\n- [x] check:harness-dirs / audit:coverage / kg:audit read storage\n\nProposal: cat-harness/docs/proposals/state-branch-2026-10-02.md\n\n## Measured 2026-10-04 (PR #2052): what the mount covers, and th", "file": "beans/defs/folio-assistant-0tg5--state-branch-p4-bean-gates-run-on-push-to-state-ha.md" }, { "id": "folio-assistant-0utt", "title": "PRODUCER: kg-export.bootstrap.qa-results.json is read by check:published-instance-exports but produced by no workflow", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T09:30:19Z", "updatedAt": "2026-10-01T18:34:24Z", "preview": "Found by the r7v6 agent (C2 fix, 2026-10-01). The C2 gate now actually compares the committed `cat-harness/test/results/kg-export.bootstrap.qa-results.json` with a fresh re-export. On its first run it found the file stale (different producer hash, untagged modules 0→3), and the file was refreshed by hand in `ba1908fa`.\n\n**Owner ruling 2026-10-01:** keep the file and add a producer. Do not delete i", "file": "beans/defs/folio-assistant-0utt--producer-kg-exportbootstrapqa-resultsjson-is-read.md" }, { "id": "folio-assistant-0uu2", "title": "DECLARATION GAP: every top-level directory is a declared subgraph — the split, and the headline count was wrong three ways", "status": "in-progress", "type": "feature", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T22:58:39Z", "updatedAt": "2026-09-24T06:22:01Z", "preview": "Owner, 2026-09-23: 'bootstrap/ should be its own subgraph.... in fact QA: in harness `/*.*` should be a subgraph.' The FIRST half only — which top-level directories are undeclared, and what each should become. The reference-direction half is a sibling session's.\n\nDeliverable is a SPLIT (declare / fold / remove / already-declared / not-mine), not a sweep. Implement only what the owner picks.\n", "file": "beans/defs/folio-assistant-0uu2--declaration-gap-every-top-level-directory-is-a-dec.md" }, { "id": "folio-assistant-0w7q", "title": "VIEWER RAIL: avatar header, per-visualiser section, drop redundant ☰/[x], QA flags (#1757)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-p5wm", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T07:25:35Z", "updatedAt": "2026-10-04T19:56:18Z", "preview": "Owner report 2026-10-01, issue #1757 (session_01Cw8JgZEDT5VqQ5ergjdMjB).\n\n## Done when\n- [x] rail header is the instance avatar or initial, clickable pin toggle; ☰ and [x] removed (rail + Jekyll footer)\n- [x] NavbarModel carries a per-visualiser section, open by default; Graphs collapsed when present\n- [x] ~~/todos/ visualiser section: todos grouped by attached KG node~~ — **struck by the owner, 2", "file": "beans/defs/folio-assistant-0w7q--viewer-rail-avatar-header-per-visualiser-section-d.md" }, { "id": "folio-assistant-0xfe", "title": "GATE RED ON MAIN: 25 published translations carry no `.po` catalogue, compounding across three merges", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T06:44:45Z", "updatedAt": "2026-09-26T06:47:24Z", "preview": "`translation-drift`'s ratchet test — \"no NEW drift, and nothing unreadable\" —\nwas failing on `main`: 25 published translations served with no `.po` catalogue\nand no entry in `UNCATALOGED`.\n\n**Fixed on `main` by #1364, not by this bean.** Closed on evidence rather than\nauthorship, per `bean-coordination` §\"Closing a bean whose work has already\nlanded\" — the evidence is re-derived below rather than ", "file": "beans/defs/folio-assistant-0xfe--gate-red-on-main-25-published-translations-carry-n.md" }, { "id": "folio-assistant-0ytk", "title": "CODE LISTS: closed sets of codes as SKOS nodes, each code with a definition and a source", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T19:29:49Z", "updatedAt": "2026-09-25T16:37:09Z", "preview": "Owner, 2026-09-23, on the adjudication answers bean bvuk had written into diagrams as bare strings: \"we need an expandable option, not just declared in code. list of codes and corresponding narrative desc and source should be part of a node/asset. can we use existing tools for new Skills?\" Chosen: all of adjudication answers, JSON-LD namespaces (2j09), our own namespaces; published through the exi", "file": "beans/defs/folio-assistant-0ytk--code-lists-closed-sets-of-codes-as-skos-nodes-each.md" }, { "id": "folio-assistant-10uc", "title": "RENDERED SURFACE: navbar, visualisers, stickies — GOAL 2's stream, moved under GOAL 4 at GOAL 2's sign-off", "status": "in-progress", "type": "task", "priority": "high", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T18:09:11Z", "updatedAt": "2026-10-16T10:30:00Z", "preview": "## What this is\n\nStream 2 of 3 in the 2026-09-22 consolidation (see the sibling claims under\n`vuip` and `yg29` for the stall measurement that prompted it). It owns **GOAL 2\n(`p5wm`) — 39 open beans**, the largest single group in the store.\n\nThis is the stream where the artefact a reader judges actually appears: the LHS\nnavbar, the visualisers, the stickies, avatars and themes.\n\n## What this stream", "file": "beans/defs/folio-assistant-10uc--stream-23-goal-2-the-rendered-surface-navbar-visua.md" }, { "id": "folio-assistant-129h", "title": "Merge refused: #2080 owed CI not green on its head", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-3fva", "blocking": [ "folio-assistant-5hox" ], "blockedBy": [], "createdAt": "2026-10-04T14:30:42Z", "updatedAt": "2026-10-07T14:55:47Z", "preview": "PR #2080, handed back by the Merge Manager on 2026-10-04 (BACKFILLED at 14:50Z: the hand-back was made as a PR comment and a queue entry only, without this bean. Owner: \"Backfill + enforce\").\n\n## Refused\n- `merge:steward` at 14:30Z: Rule_HeadNotGreen: conflict declared, own CI missing-required.\n- Queue entry (`beans/queue/litlfred--folio-assistant--2080.json`, status `waiting-on-author`): Owner-ap", "file": "beans/defs/folio-assistant-129h--merge-refused-2080-owed-ci-not-green-on-its-head.md" }, { "id": "folio-assistant-12s9", "title": "PREFIX = DECLARING SUBGRAPH: rename the folio: BPMN prefix and folio-*/v1 schema ids to the path of the Subgraph that declares them (owner ruling iwtn #1)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-88mg", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T21:03:10Z", "updatedAt": "2026-09-24T18:02:28Z", "preview": "## Owner ruling, 2026-09-23 (bean iwtn, ruling 1 of 4)\n\n> use `:` as prefix when assets are declared/instantiated in ``. what patterns for sub-sub-graphs? `:::` or so?\n\nA nested Subgraph uses its **path of Subgraph ids joined with dots**. XML forbids a colon inside a prefix, so `a:b:c:` cannot be used. The owner chose this pattern on 2026-09-23 over \"Harness name only\" a", "file": "beans/defs/folio-assistant-12s9--prefix-declaring-subgraph-rename-the-folio-bpmn-pr.md" }, { "id": "folio-assistant-12ws", "title": "ONE NAME, TWO MEANINGS: check-l1-complete declares a local instanceRootFor that contradicts the exported one", "status": "completed", "type": "task", "priority": "low", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-21T07:36:38Z", "updatedAt": "2026-10-07T06:34:00Z", "preview": "FOUND 2026-09-21 while sweeping `a6kl`.\n\n`cat-harness/schemas/cat-harness.ts:2495` exports:\n\n /** {@link findInstanceRoot}, throwing rather than returning `undefined`. */\n export function instanceRootFor(start: string): string\n\n`cat-harness/scripts/check-l1-complete.ts:718` declares its own:\n\n export function instanceRootFor(cwd: string): string | undefined\n\nThey are genuinely different f", "file": "beans/defs/folio-assistant-12ws--one-name-two-meanings-check-l1-complete-declares-a.md" }, { "id": "folio-assistant-1335", "title": "DAK block kinds into smart-base: a harness contributes its content adapter and block kinds to core (dak-blocks.ts, DAK entries of block-kinds.ts)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-n3ni", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T17:45:13Z", "updatedAt": "2026-10-02T15:26:21Z", "preview": "Split off stage D (bean kg83) of the smart-* separation (#1767), owner 2026-10-01: 'Split: move 2 now, bean the rest'.\n\nStage D moved dak.ts and dak-content-type.ts (and their tests) into smart-base/schemas/. dak-blocks.ts could NOT follow, measured 2026-10-01: core's own content model knows DAK —\n- cat-harness/schemas/block-kinds.ts: CONTENT_ADAPTERS = ['paper', 'dak'], DAK_BLOCK_KINDS, DAK_COMPO", "file": "beans/defs/folio-assistant-1335--dak-block-kinds-into-smart-base-a-harness-contribu.md" }, { "id": "folio-assistant-14ve", "title": "REGEN NOT A FIXPOINT: bun run cat regen asks each check once, so a check asked before its input's writer runs reports current and stays stale", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-29T23:42:28Z", "updatedAt": "2026-10-08T05:42:00Z", "preview": "Measured 2026-09-29 on claude/sharp-einstein-970n6g (PR #1511), after merging main:\n\n- `bun run cat regen` pass 1: \"58 current, 4 regenerated, 0 unrepaired\", exit 0.\n- `navbar:include:check` still failed, and so did `gen-navbar-include.test.ts` (\"the writer is a fixpoint over the real data\").\n- Cause: the navbar include is generated from `cat-harness/docs/_data/harness.json`, which `docs:harness` ", "file": "beans/defs/folio-assistant-14ve--regen-not-a-fixpoint-bun-run-regen-asks-each-check.md" }, { "id": "folio-assistant-16ei", "title": "qa-store: one read/write API for QA results, qa:fetch / qa:publish, ContentDirectory.storage, CI publish and a prune that fires", "status": "completed", "type": "feature", "priority": "high", "parent": "folio-assistant-3fva", "blocking": [ "folio-assistant-2ae2", "folio-assistant-c8uq", "folio-assistant-cxcn", "folio-assistant-d6bw", "folio-assistant-id4s" ], "blockedBy": [], "createdAt": "2026-10-01T08:00:46Z", "updatedAt": "2026-10-06T06:09:21Z", "preview": "Arc `3fva`, proposal §2.2–2.4 and §4 Phase 2. Blocked on the SPIKE bean. Serial, one agent: everything else depends on this.\n\n- `qa-store` module:\n - `readQa(ref, path)` returns hit / miss / corrupt / unknown, and a miss is never read as clean;\n - `publishQa(ref)` follows the lake-cache write path, with fetch → rebuild-on-tip → push, 3 attempts and `backoff-sleep.ts`, and never `-f`.\n- `bun run ", "file": "beans/defs/folio-assistant-16ei--qa-store-one-readwrite-api-for-qa-results-qafetch.md" }, { "id": "folio-assistant-18p1", "title": "EXPORT test-plan/v1 → FHIR R5 TestPlan in fhir-harness, once SMART Guidelines IG content is complete", "status": "draft", "type": "feature", "priority": "normal", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [ "folio-assistant-ygzh" ], "createdAt": "2026-10-01T09:12:25Z", "updatedAt": "2026-10-01T09:12:25Z", "preview": "Owner, 2026-10-01: *\"https://www.hl7.org/fhir/testplan.html is a downstream target in fhir context … we should be able to map to fhir test plan when we have full smart guidelines IG content.\"*\n\n## What\n`test-plan/v1` (platform, `cat-harness/schemas/test-plan.ts`, bean `ygzh`) stays FHIR-free. In the FHIR context a SMART Guidelines IG gets its test plans **exported as FHIR TestPlan resources**:\n- `", "file": "beans/defs/folio-assistant-18p1--export-test-planv1-fhir-r5-testplan-in-fhir-harnes.md" }, { "id": "folio-assistant-1948", "title": "SHA-pin GitHub Actions workflows: required on main and publish, unpinned on staging", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-ieum", "blocking": [], "blockedBy": [], "createdAt": "2026-10-07T22:03:25Z", "updatedAt": "2026-10-08T03:00:00Z", "preview": "## Landed evidence\n- Completed as part of epic `folio-assistant-ieum` on main (PR #2504).\n- SHA-pinning required on main and publish workflows, verified on origin/main.", "file": "beans/defs/folio-assistant-1948--sha-pin-github-actions-workflows-required-on-main.md" }, { "id": "folio-assistant-197s", "title": "Dependabot cannot run Bun here: 'could not run Bun … configuration error' on #908, and the npm ecosystem never touches bun.lock", "status": "todo", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T16:58:15Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "Measured 2026-09-26.\n\n## What happened\n\n- At 15:26Z dependabot commented on #908: *\"Dependabot could not run Bun to update your dependencies due to a configuration error.\"* That came about 13 minutes after #1407 put the same bumps on main.\n- **This is not evidence of a defect in our config.** The same message appeared on 25–26 Sep across many unrelated repos with different setups. Examples: [bun-c", "file": "beans/defs/folio-assistant-197s--dependabot-cannot-run-bun-here-could-not-run-bun-c.md" }, { "id": "folio-assistant-19wc", "title": "Skill pr-description: a PR body a first-time reader can follow", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T15:59:38Z", "updatedAt": "2026-10-02T20:46:38Z", "preview": "Owner, 2026-10-02 (on litlfred/fhir-ig-publisher#8): *\"PR should be readable for human for first time. what probelm trying to address, why, how approached, constratins, soltion... ...usage\"* and *\"make that a general skill when writing/updating PRs. see existing skills/processes\"*. Issue #1879.\n\n## Done when\n- [ ] cat-harness/skills/sdlc/sdlc-core/pr-description.md written and registered (skill:re", "file": "beans/defs/folio-assistant-19wc--skill-pr-description-a-pr-body-a-first-time-reader.md" }, { "id": "folio-assistant-1acg", "title": "check:head-has-run counts an action_required run as RAN — a run that never executed satisfies 'the gates fired'", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T00:15:20Z", "updatedAt": "2026-10-03T01:21:22Z", "preview": "## The defect, in one line\n\n`ranAs` matches a run on **name and event** and never reads its `conclusion`:\n\n```ts\nfunction ranAs(runs: RunRow[], t: WorkflowTrigger, event: string): boolean {\n return runs.some((r) => r.name === t.name && r.event === event);\n}\n```\n\n`RunRow` already carries `conclusion`. The data was in hand and not consulted.\n\nA `pull_request` run that completes **`action_required`*", "file": "beans/defs/folio-assistant-1acg--checkhead-has-run-counts-an-action-required-run-as.md" }, { "id": "folio-assistant-1br0", "title": "SEARCH REMOTE: publish the identifier lookup as its own page and link it from the search box (#1972 step 3)", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-whlc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T14:17:59Z", "updatedAt": "2026-10-03T14:17:59Z", "preview": "Issue #1972 step 3 (remote/large graphs). The owner's choice of 2026-10-03,\nfrom four options: **a separate page and a link** — publish the identifier\nlookup as its own page OUTSIDE the docs pipeline, name it in the search\nmanifest as a `remote` scope, and have the search box show one link to it.\n\nIt keeps both standing decisions: bean `4pm8`'s ruling (title search over\nreferenced nodes is an acce", "file": "beans/defs/folio-assistant-1br0--search-remote-publish-the-identifier-lookup-as-its.md" }, { "id": "folio-assistant-1bvx", "title": "check:reference-direction fails on main and is in NO workflow — nothing catches it", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T22:13:49Z", "updatedAt": "2026-10-06T14:30:00Z", "preview": "Measured on `main` `25bff68c4f3`, 2026-09-30.\n\n```\n$ bun run cat check:reference-direction\n✗ 1 PENDING entr(y/ies) no longer qualify — delete them\n✗ 91 file(s) name several instances above them and are not in PENDING\nexit 1\n```\n\n```\n$ grep -q \"bun run cat check:reference-direction\" .github/workflows/*.yml\nnot in any workflow\n```\n\n**Red on the default branch, and run by nothing.** That is `1xhc` in", "file": "beans/defs/folio-assistant-1bvx--checkreference-direction-fails-on-main-and-is-in-n.md" }, { "id": "folio-assistant-1dre", "title": "PAGES SERVING PROBE: the instrument pages-publish-health counts, and it can only be built in CI because egress refuses every route from an agent", "status": "completed", "type": "feature", "priority": "", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T03:55:45Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "`pages-publish-health` (bean `qj9a`) counts instruments for observing the\npublished site. The count is **0**, so it reports a `major` finding whose action\nis this bean. This is that probe, specified rather than built, because it cannot\nbe built here.\n\n## Why it cannot be written or tested from an agent's container\n\nMeasured 2026-09-25 and 2026-09-26, every route refused:\n\n| route | result |\n|---|-", "file": "beans/defs/folio-assistant-1dre--pages-serving-probe-the-instrument-pages-publish-h.md" }, { "id": "folio-assistant-1g4s", "title": "OPTION A: subgraphs inherit automatically as named members; dependents retired", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-iirv", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T22:39:25Z", "updatedAt": "2026-10-01T06:58:18Z", "preview": "Owner 2026-09-30: 'if f-a-core ... has a docs/ ... and depends on cat-harness where docs/ is declared ... the viewer should detect f-a-core/docs ... a named (sub-)subgraph of docs' and 'make dependents: reproduce automatic behaviour so don't need it'.\n\n## Summary of Changes\n- resolveDirectories: every instance-scoped entry gains one MEMBER per chain instance whose / exists (ResolvedDir", "file": "beans/defs/folio-assistant-1g4s--option-a-subgraphs-inherit-automatically-as-named.md" }, { "id": "folio-assistant-1gf7", "title": "ASSESS: open standards for requirements, and methodologies that fit the bootstrap Requirement", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-2upx", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T19:34:01Z", "updatedAt": "2026-10-09T18:43:00Z", "preview": "Issue #1164. Owner, 2026-09-23, choosing markdown front matter for requirement documents: '1 for now. need bean to assess open stds for requirements/methodologies that fit in'.\n\nThe bootstrap Requirement is deliberately free of outside concepts. This bean asks which open standards a harness could MAP it to or from, in a layer above bootstrap, and at what cost.\n\n## Done when\n- [x] a survey of open ", "file": "beans/defs/folio-assistant-1gf7--assess-open-standards-for-requirements-and-methodo.md" }, { "id": "folio-assistant-1h6u", "title": "main red after 2c295f8 + translation batches #1404/#1409: 6 dangling skill links, 9 stale derived sets", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T14:30:05Z", "updatedAt": "2026-09-26T14:34:12Z", "preview": "Measured on pristine main f850f721a06, 2026-09-26. The failing checks were glossary-page, gen-skill-docs, gen-docs-pages, docs:auto, translation:index, check:ci-invocations (which re-runs the two generators), check:subgraphs, state:visualizer and docs:harness. Every open PR inherited them through its merge ref.\n\nCauses:\n- **2c295f8**, committed straight to main, rewrote skill script paths as `../.", "file": "beans/defs/folio-assistant-1h6u--main-red-after-2c295f8-translation-batches-1404140.md" }, { "id": "folio-assistant-1hjm", "title": "merge queue is UNAVAILABLE on this repo — three workflows carry merge_group triggers that can never fire", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T22:29:45Z", "updatedAt": "2026-10-07T04:47:41Z", "preview": "Measured 2026-09-30, after the owner asked for the merge queue to be turned on.\n\n## What was found\n\nThe repository's **New branch ruleset** form offers thirteen rules. **`Require\nmerge queue` is not one of them**:\n\n```\nRestrict creations · Restrict updates · Restrict deletions\nRequire linear history · Require deployments to succeed · Require signed commits\nRequire a pull request before merging · R", "file": "beans/defs/folio-assistant-1hjm--merge-queue-is-unavailable-on-this-repo-three-work.md" }, { "id": "folio-assistant-1j3q", "title": "ROUTE-KEYED STORAGE: a third `keyedBy` for regenerable rendered pages — one entry per route, replaced by its one writer, never spliced with `expect`", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [ "folio-assistant-xsrv" ], "blockedBy": [], "createdAt": "2026-10-03T13:20:54Z", "updatedAt": "2026-10-03T15:15:19Z", "preview": "Owner, 2026-10-03, decision 3 of the auto-doc rendered-graphs Q&A (#1966):\n**\"Add route-keyed storage, then move it off main\"** — asked about the cost of\nleaving generated pages committed on `main`.\n\nThe schema's own docblock (`cat-harness.ts`, §\"`keyedBy` — two keyings, and a\nthird is a schema change\") says what this bean is: *\"The field is an enum, not\na string, so a third keying is a schema cha", "file": "beans/defs/folio-assistant-1j3q--route-keyed-storage-a-third-keyedby-for-regenerabl.md" }, { "id": "folio-assistant-1l13", "title": "witness-pipeline.yml has never run — keep, template, or retire?", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T06:47:37Z", "updatedAt": "2026-10-23T08:00:00Z", "preview": "Split from 52dz (2026-09-24). 52dz's owner decisions moved section-title-audit into folio_init templates, but witness-pipeline.yml was not part of that decision and still sits in .github/workflows with ZERO runs all time (measured 2026-09-20), so check:ci-health cannot see it by construction.\n\n## Done when\n- [x] measured: does it run against the platform at all, or only a folio's computations? (me", "file": "beans/defs/folio-assistant-1l13--witness-pipelineyml-has-never-run-keep-template-or.md" }, { "id": "folio-assistant-1lfx", "title": "DEPLOY: STAGING must report which host rendered it, not assume gh-pages", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T08:55:36Z", "updatedAt": "2026-10-09T14:05:00Z", "preview": "From [#363](https://github.com/litlfred/folio-assistant/issues/363): \"make sure\nSTAGING knows where rendering happes (ghpages, local server) so can tell user\ncorrectly\".\n\n## The defect, stated as what goes wrong\n\nAn agent finishing a change tells the author where to look. Today that sentence\nis composed on the assumption that rendering went to GitHub Pages. On a private\nrepository there is no Page", "file": "beans/defs/folio-assistant-1lfx--deploy-staging-must-report-which-host-rendered-it.md" }, { "id": "folio-assistant-1nu0", "title": "Viewers scoped to cat-harness only: Tools, QA, Beans, Health, Folio, Glossary, Attestations, Schemas", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-46ob", "blocking": [], "blockedBy": [], "createdAt": "2026-10-15T16:53:28Z", "updatedAt": "2026-10-15T18:59:39Z", "preview": "## Parent\n\nfolio-assistant-46ob\n\nA viewer exists, but it is declared with `covers: []` (cat-harness's own directory) or per-instance sub-graphs that were never drawn, so other harnesses' rows read \"no viewer yet\".\n\nMeasured live: Tools ×6 (cat-harness `tools` visualiser, covers id `tools`; page lists cat-harness's tools only), QA ×5 (`lsi`, covers id `qa`), Health ×3, Beans ×2, Folio", "file": "beans/defs/folio-assistant-1nu0--viewers-scoped-to-cat-harness-only-tools-qa-beans.md" }, { "id": "folio-assistant-1ple", "title": "state:mount follows an ABSOLUTE marker path out of its own checkout: a copied checkout re-mounts (and can empty) the original", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-14T13:19:55Z", "updatedAt": "2026-10-16T07:48:00Z", "preview": "Found 2026-10-14 (session_01XueePufkpx8fwJxavyDS5h; owner chose \"C1: track it as a bean\").\n\n## What happened\nA scratch copy of an index checkout (`cp -a /home/user/folio-assistant /S`) carried the original's `.git/branch-mounts/*.json` markers. Each marker's `into` is an ABSOLUTE path (`/home/user/folio-assistant/beans`, `…/todos`, `…/fsh-guts`, `…/issue-marks`). `bun run cat state:mount`", "file": "beans/defs/folio-assistant-1ple--state-mount-follows-an-absolute-marker-path-out-of.md" }, { "id": "folio-assistant-1q4b", "title": "Re-draw the navbar wireframe after ob3m's 12 findings landed", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T05:55:48Z", "updatedAt": "2026-10-06T18:30:00Z", "preview": "ob3m's closing step, split out when ob3m closed on 2026-10-06: 'When fixed, re-draw cat-harness/docs/wireframes/navbar/ and re-run wireframe:check and check:wireframes.' intent.md's Observed and Findings sections still describe main with #1010/#1022 (before #1762, #1804, #1805, #1807, #1808 and #1819), so the wireframe documents a navbar that no longer exists.\n\n## Done when\n- [x] as-is.html and in", "file": "beans/defs/folio-assistant-1q4b--re-draw-the-navbar-wireframe-after-ob3ms-12-findin.md" }, { "id": "folio-assistant-1qrr", "title": "FEATURE-STAGING: the workflow definition comes from the BASE but the checkout is the PR HEAD, so a newly-added step fails every branch that predates it", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T15:33:32Z", "updatedAt": "2026-10-22T17:10:00Z", "preview": "## The split\n\n`.github/workflows/feature-staging.yml:190` checks out the PR **head**:\n\n```yaml\nref: ${{ github.event.pull_request.head.ref || inputs.branch || github.ref }}\n```\n\nBut for a `pull_request` event GitHub takes the workflow **definition** from the\nBASE branch. So the steps that run are `main`'s, while the tree they run\nagainst is the PR's head. Those two disagree the moment a step is ad", "file": "beans/defs/folio-assistant-1qrr--feature-staging-the-workflow-definition-comes-from.md" }, { "id": "folio-assistant-1r0p", "title": "INGEST: audio — transcription and translation", "status": "todo", "type": "task", "priority": "normal", "parent": "folio-assistant-slw1", "blocking": [], "blockedBy": [], "createdAt": "2026-09-16T06:43:50Z", "updatedAt": "2026-10-21T21:32:00Z", "preview": "## What\n\nAudio content is transcribed, and the transcript translated into the folio's\nlanguages.\n\n## Done when\n\n`library//transcript/` holds the source-language transcript and each\ntranslation, both referenced from `manifest.jsonld`, both carrying the\nprovenance stamp (`folio-assistant-iqim`), and both subject to the round-trip\nQA (`folio-assistant-ktt2`).\n\nDiagram: `processes/ingest-derive-", "file": "beans/defs/folio-assistant-1r0p--ingest-audio-transcription-and-translation.md" }, { "id": "folio-assistant-1s5s", "title": "The published package's PYTHON half runs nowhere — CI's pytest job globs a directory that does not contain it", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T09:13:49Z", "updatedAt": "2026-09-26T10:38:03Z", "preview": "Split out of `rsi6` 2026-09-26. **That bean was closed `completed` carrying two\nunchecked boxes**, which the store's own rule forbids — \"ONLY if the bean has\nno unchecked todo items left\". Recorded here rather than quietly re-ticked,\nbecause a completed bean holding live work is exactly the shape that makes the\nstore unreadable, and this session has spent the day filing that class against\nother pe", "file": "beans/defs/folio-assistant-1s5s--the-published-packages-python-half-runs-nowhere-ci.md" }, { "id": "folio-assistant-1swy", "title": "QA: verdicts, sidecars and the audited review record", "status": "in-progress", "type": "epic", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T11:43:44Z", "updatedAt": "2026-10-04T15:12:14Z", "preview": "Where a verdict lives, what shape it has, and who is allowed to change it.\n\n`2634` is the placement rule (provenance decides: a QA reviewer's output lives\nunder `test/results/`), `9gyz` is the shape (Finding, Decision and AuditNote are\nthree entities), and `cflw`/`nytj` are the two ways concurrent work corrupts the\nrecord — 218 sidecars rewritten by one edit, and two PRs green alone but red\ntogeth", "file": "beans/defs/folio-assistant-1swy--qa-verdicts-sidecars-and-the-audited-review-record.md" }, { "id": "folio-assistant-1xhc", "title": "CI RELIABILITY: a gate that does not fire is indistinguishable from one that passed", "status": "in-progress", "type": "epic", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T11:43:44Z", "updatedAt": "2026-10-04T15:12:14Z", "preview": "A gate that does not fire is indistinguishable from one that passed.\n\nThat is `xom7`'s lesson one layer up, and every item here is an instance:\n`3pqn` (a PR that opens with zero checks looks exactly like one whose checks\nare green), `dzl3` (playwright fails before any test body runs), `t373`\n(`readme:audit` is in no workflow at all, and `main` carried eight dead links),\n`bgle` (a guard that enumer", "file": "beans/defs/folio-assistant-1xhc--ci-reliability-a-gate-that-does-not-fire-is-indist.md" }, { "id": "folio-assistant-1xrg", "title": "state:visualizer:check cannot fail on a declaration that points at its own page (the '4 of 7 render a shell' premise was a FALSE FINDING — retracted)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T06:48:40Z", "updatedAt": "2026-09-30T09:03:42Z", "preview": "Found 2026-09-22 by LOOKING at a page rather than at the generator that wrote\nit. `lqo9` slice 2 added the `glossary` state graph, `state:visualizer`\ngenerated `docs/glossary/index.html` for it, every gate went green, and the\npage displays **nothing about the glossary**.\n\nIt is not a defect that slice introduced. Measured across all seven state\ngraphs the visualizer writes a page for:\n\n| graph | p", "file": "beans/defs/folio-assistant-1xrg--state-dashboards-4-of-7-render-a-shell-a-declared.md" }, { "id": "folio-assistant-1yd7", "title": "LANDING STICKIES: each instance's sticky lives in its own folio/, not in cat-harness/folio/", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-7x5n", "blocking": [], "blockedBy": [], "createdAt": "2026-10-07T21:46:20Z", "updatedAt": "2026-10-09T14:05:00Z", "preview": "## What\n\n`cat-harness/folio/` holds the landing stickies (`folio-landing-sticky/v1`) for\nother instances, beside its own `cat-harness.json`:\n`who-iris.json`, `bootstrap.json`, `folio-assist-core.json` and `folio-assistant.json`.\nMeasured on main 33bf3d2, 2026-10-07.\n\nEach sticky belongs in the `folio/` directory of the instance it describes.\n\n## Why now\n\nThe separation (ar1s) moves instances to th", "file": "beans/defs/folio-assistant-1yd7--landing-stickies-live-with-their-own-instance-not-in-cat-h.md" }, { "id": "folio-assistant-1ygp", "title": "ROAST: adversarial pass over the hand-over screen, its wiring and the zero-trust-handover methodology before adoption", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ieum", "blocking": [], "blockedBy": [], "createdAt": "2026-10-07T19:29:09Z", "updatedAt": "2026-10-08T00:15:43Z", "preview": "Owner 2026-10-07: 'include adversarial analysis (general process we have)'. The process: lens fan-out plus an adjudicator (devils-advocate-watcher, Slot C), then generalise-the-fix move 2 on the wiring. Findings are recorded HERE with a status (answered / open / accepted-as-cost), because a printed objection is gone (v048).\n\n## Done when\n- [x] lenses run, adjudicated\n- [x] every surviving finding ", "file": "beans/defs/folio-assistant-1ygp--roast-adversarial-pass-over-the-hand-over-screen-i.md" }, { "id": "folio-assistant-24fa", "title": "beans:claim pushes to main, which GUARANTEES a refuse-class conflict on the claiming branch's own completion", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-d33q", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T00:43:10Z", "updatedAt": "2026-10-07T17:38:00Z", "preview": "## Measured on itself, 2026-10-03\n\n`bun run cat beans:claim 8rff` writes the claim **straight to `main`**:\n\n> `_2026-10-03T00:27:48Z_ — Claimed by claude/merge-patterns-8rff — pushed to\n> main so sibling sessions see it before this branch has a PR (bean 35nj)`\n\nThat is deliberate and it works — `35nj` (\"A bean claim is invisible to a\nsibling until the PR exists, so claim-before-work does not preve", "file": "beans/defs/folio-assistant-24fa--beansclaim-pushes-to-main-which-guarantees-a-refus.md" }, { "id": "folio-assistant-2634", "title": "QA outputs live under test/results/ — verdicts and witnesses both, by provenance", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T07:11:55Z", "updatedAt": "2026-10-07T17:32:00Z", "preview": "THE RULE, from the owner 2026-09-19: 'if the witnesses were generated as a QA reviewer primarily then it should be under test/results/ as part of a QA process.'\n\nPlacement follows PROVENANCE, not file family and not consumption. An artefact that is primarily the output of a QA review belongs under `test/results/`, whatever fetches it afterwards.\n\n## What exists today, measured 2026-09-19 on main\n\n", "file": "beans/defs/folio-assistant-2634--qa-outputs-live-under-testresults-verdicts-and-wit.md" }, { "id": "folio-assistant-269z", "title": "Folio handle covers the who-iris replica's INGESTED COPY banner — conflicts with jpjt 'replica unchanged'", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T18:04:45Z", "updatedAt": "2026-09-24T18:04:45Z", "preview": "Found by bean 015u (2026-09-24). On the who-iris replica the fixed ▾ Folio handle covers the 'INGESTED COPY — not WHO' banner at 390 px, and at 1280 px a link sits centred under it, so a click there opens the glass instead.\n\nThe obvious fix, reserving the handle's band with top padding, is ruled out by an existing rule: `folio-mount.e2e` asserts a replica is UNCHANGED while the glass is closed (be", "file": "beans/defs/folio-assistant-269z--folio-handle-covers-the-who-iris-replicas-ingested.md" }, { "id": "folio-assistant-279l", "title": "B10b (#1168): glossary ledger keys lanes by BPMN id; SKOS altLabel for renames; key pattern enforced + QA gate", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T14:59:53Z", "updatedAt": "2026-09-30T17:07:28Z", "preview": "Owner 2026-09-30: do as rec + QA gate + use SKOS for rename/alternate name. lane/ → process/

/lane/ (kg-export's identity); propertyNames pattern; a gate that fails on an off-pattern key; a renamed lane keeps its key and the old name becomes skos:altLabel (not a new term).\n## Done when\n- [ ] ledger keys match ^(role|process)/…; lane/Actor migrated\n- [ ] rename handled as altLabel/h", "file": "beans/defs/folio-assistant-279l--b10b-1168-glossary-ledger-keys-lanes-by-bpmn-id-sk.md" }, { "id": "folio-assistant-27u0", "title": "check:anchor-names was blind to 2 of 4 base spellings: 75 ascents unseen, 9 real findings, and it printed a green check", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-27T05:47:27Z", "updatedAt": "2026-09-27T05:47:27Z", "preview": "Found 2026-09-27 while reading `check-translation-catalogue.ts`, which arrived on\n`main` in the 36 commits merged into `claude/brave-hypatia-r820sf`. Its docblock\njustified deriving the instance root locally with *\"because no module exports\nit\"*. That claim is false — `cat-harness.ts:2702` exports `instanceRootFor` — and\nchecking it opened the real defect one level down.\n\n## The defect\n\n`check:anc", "file": "beans/defs/folio-assistant-27u0--checkanchor-names-was-blind-to-2-of-4-base-spellin.md" }, { "id": "folio-assistant-2ae2", "title": "Readers: docs site, feature staging, review heat map and MCP tools fetch QA from qa-reports", "status": "scrapped", "type": "task", "priority": "normal", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [ "folio-assistant-16ei" ], "createdAt": "2026-10-01T08:00:46Z", "updatedAt": "2026-10-10T17:04:26Z", "preview": "Arc `3fva`, proposal §4 items 3.3 and 3.4. Blocked on the qa-store bean. Can be dispatched in parallel with the gates bean.\n\n- the `assets/qa/` copy in `docs-site.yml:609` and `feature-staging.yml:1089`\n- the badges (`head_custom.html`, `docs/qa/index.html`)\n- `publish-block-qa.ts`, `review-heat.ts` and `gen-review-page.ts`\n- MCP: lsi output (`tools/index.ts`), `src/tools/degradation.ts` and `src/", "file": "beans/defs/folio-assistant-2ae2--readers-docs-site-feature-staging-review-heat-map.md" }, { "id": "folio-assistant-2as0", "title": "Agent-triggered staging: local build + push to gh-pages, then an Artifact preview (#2410)", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-q4jm", "blocking": [], "blockedBy": [], "createdAt": "2026-10-07T11:12:53Z", "updatedAt": "2026-10-07T12:44:42Z", "preview": "Issue #2410. PR 1: cat-harness/scripts/stage-local.ts builds the folio's preview in the agent's checkout and pushes STAGING// through the same gate and render log as the workflow. PR 2: --artifact publishes it as a claude.ai Artifact. Measured 2026-10-07: 33 min runner queue before the workflow ran 45 s; Pages deploy 47 min queued.\n\n*2026-10-07* — completed: PR 1 (stage-local, push to gh-pag", "file": "beans/defs/folio-assistant-2as0--agent-triggered-staging-local-build-push-to-gh-pag.md" }, { "id": "folio-assistant-2b5s", "title": "MOUNT COPIES A DIRECTORY WHOLESALE: /who-iris/ publishes 1,367 corpus files as pages, and /library/who-iris/ publishes all 1,378 a second time", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-yj32", "blocking": [], "blockedBy": [], "createdAt": "2026-09-21T20:00:52Z", "updatedAt": "2026-09-30T13:57:36Z", "preview": "## What\n\n`mount-instance-docs.ts` mounts a directory when it carries an `index.html`\n**at its own root**, and then copies **everything beneath it**. For\n`who-iris/library/` that is 1,378 files, of which 11 are pages:\n\n| | count | what |\n|---|---|---|\n| pages | 11 | 7 `.html` (index, community-list, 2 collections, 3 items), 3 cover `.png`, `image-verdicts.json` |\n| L1 corpus | 1,367 | 813 `.jsonld`", "file": "beans/defs/folio-assistant-2b5s--mount-copies-a-directory-wholesale-who-iris-publis.md" }, { "id": "folio-assistant-2c2b", "title": "CI WATCH IS TWO-STATE: cancelled read as failure and reported a green main red — bun run cat ci:watch, three states with undetermined never collapsing", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T14:56:30Z", "updatedAt": "2026-09-30T19:43:02Z", "preview": "Issue #1624. Owner asked for this as a repo tool rather than session-local\nscratch tooling, 2026-09-30.\n\n## The defect, measured\n\nAn agent's CI poll folded `cancelled` into the failure set and reported **`main`\nred** on `3d2c7fbe869`, whose every hard gate had passed:\n\n```\nsuccess Repository gates (hard)\nsuccess TypeScript — tests, lint, types (hard)\nsuccess End-to-end + accessibility (hard)", "file": "beans/defs/folio-assistant-2c2b--ci-watch-is-two-state-cancelled-read-as-failure-an.md" }, { "id": "folio-assistant-2coi", "title": "IG fork sites: artefact JSON, JSON Schema and JSON-LD from the AST route, with the AST cache built and seeded in CI", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-11T07:17:27Z", "updatedAt": "2026-10-11T12:25:41Z", "preview": "Owner, 2026-10-11: \"fix: Artefact pages still link JSON to WHO, and have no JSON-LD or JSON Schema links. Use AST route. make sure it is fullly working w/ branch caches etc\". Follows 48a6 (fork site drift).\n\nLanded: fhir-harness#38 (ast-to-artifact-index --sidecars; ig-cache keeps expansions.json), #39 (--keep-published-grouping; description links rebased); smart-base#42 (smart-base:dak-ast-sideca", "file": "beans/defs/folio-assistant-2coi--ig-fork-sites-artefact-json-json-schema-and-json-l.md" }, { "id": "folio-assistant-2e39", "title": "B9b (#1168): a BPMN lane is its own node; performedBy points at the registry role", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T10:54:25Z", "updatedAt": "2026-09-30T12:36:35Z", "preview": "Owner, 2026-09-30: Lane node. kg-export mints role/ per lane, so one role can get two nodes or merge by accident. Emit process/

/lane/ (type Lane) with bindsRole; performedBy goes to role/.\n\n## Done when\n- [ ] lanes exported as Lane nodes, not Role nodes\n- [ ] performedBy resolves to registry Role nodes, 0 dangling\n\n\n\n## Done — #1587\nLane nodes (process/

/lane/, par", "file": "beans/defs/folio-assistant-2e39--b9b-1168-a-bpmn-lane-is-its-own-node-performedby-p.md" }, { "id": "folio-assistant-2ee5", "title": "Merge refused: #1955 handed back", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-whlc", "blocking": [ "folio-assistant-4ak5" ], "blockedBy": [], "createdAt": "2026-10-04T14:30:43Z", "updatedAt": "2026-10-07T14:55:47Z", "preview": "PR #1955, handed back by the Merge Manager on 2026-10-04 (BACKFILLED at 14:50Z: the hand-back was made as a PR comment and a queue entry only, without this bean. Owner: \"Backfill + enforce\").\n\n## Refused\n- `merge:steward` at 14:30Z: Rule_MvpHighRisk: conflict declared, own CI green.\n- Queue entry (`beans/queue/litlfred--folio-assistant--1955.json`, status `active`): STATUS UNKNOWN on the 13:40Z al", "file": "beans/defs/folio-assistant-2ee5--merge-refused-1955-handed-back.md" }, { "id": "folio-assistant-2gst", "title": "QA READERS F1: kg-audit reads its carried state (pair attestations, voice reviews, test runs) from qa-store — two false-cleans, and the kg-qa D2 split", "status": "completed", "type": "task", "priority": "critical", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T08:47:12Z", "updatedAt": "2026-10-23T13:35:00Z", "preview": "Arc `3fva`, from reader audit `gxvk` (`cat-harness/docs/proposals/qa-readers-audit-2026-10-01.md`, §5.1 family F1). Refines `oqe3` 3.1a. Blocked on the qa-store bean `16ei`.\n\n**Holds two of the audit's CRITICAL rows.** It must land before `5hox`: deleting the kg-qa tree before this lands destroys the attestations without a word.\n\n## Readers\n- `cat-harness/scripts/kg-audit.ts:2751`: `--check` stale", "file": "beans/defs/folio-assistant-2gst--qa-readers-f1-kg-audit-reads-its-carried-state-pai.md" }, { "id": "folio-assistant-2h76", "title": "STATE BRANCH P2: mechanism — storage keyedBy tip, branch-store splice-write, seed the orphan 'state' branch, session-start mount at state/", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T10:58:10Z", "updatedAt": "2026-10-06T06:26:35Z", "preview": "Serial, one agent. Reuses the 3fva spike 3ds9 write path (hash-object, mktree, commit-tree, push; never -f).\n\n## Done when\n- [x] schema field with keyedBy: tip\n- [x] branch-store.ts with retry over a moved tip\n- [x] state branch seeded with a hash-verified manifest (2026-10-02, see below)\n- [x] session-start hook mounts state/ and fails LOUDLY when it cannot\n- [x] bun run cat state:push\n- [x] meas", "file": "beans/defs/folio-assistant-2h76--state-branch-p2-mechanism-storage-keyedby-tip-bran.md" }, { "id": "folio-assistant-2i5f", "title": "TERMINOLOGY / adjudication: the extractor vs the terminology, and judge vs judge — call the existing process, restate nothing", "status": "in-progress", "type": "task", "priority": "normal", "parent": "folio-assistant-5yhm", "blocking": [], "blockedBy": [], "createdAt": "2026-09-25T04:51:47Z", "updatedAt": "2026-10-02T08:01:23Z", "preview": "Under `5yhm`. **Two disagreements, not one**, and only the second is already\nsolved here.\n\n1. **The extractor and the terminology disagree.** The corpus uses a word the\n authorised vocabulary spells differently, or does not carry at all. A\n *content* decision — change the prose, author a local term with its reason,\n or record that the vocabulary is wrong for this domain. Nothing covers this.", "file": "beans/defs/folio-assistant-2i5f--terminology-adjudication-the-extractor-vs-the-term.md" }, { "id": "folio-assistant-2ipj", "title": "PROCESS CATALOGUES: clinical (smart-base-clinical holds none yet), public health and health system (from the DTH documents)", "status": "todo", "type": "feature", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-10T14:06:52Z", "updatedAt": "2026-10-20T18:19:02Z", "preview": "Owner, 2026-10-10: *\"there is a WHO/smart-base-clinical ... which will be the catalog of clinical processes/encoutners. Assume that such analouges will exist for public health and health systems (these will come fromt he already ingested Digiatl Transformation Handbooks (for supply chain architecture, product catalogu,,,) and more are coming).\"*\n\n## Measured 2026-10-10\nWorldHealthOrganization/smar", "file": "beans/defs/folio-assistant-2ipj--process-catalogues-clinical-smart-base-clinical-ho.md" }, { "id": "folio-assistant-2j09", "title": "The undeclared-namespace check reads only BPMN xmlns, so a JSON-LD vocabulary can go unpinned forever", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T11:48:29Z", "updatedAt": "2026-09-23T19:29:49Z", "preview": "Split out of the 2026-09-22 fix to `external-schemas.ts`, deliberately not taken there because it is a much larger decision than the bug beside it.\n\n## The fix that was taken\n\n`unusedNamespaces` was reporting Dublin Core — 43 real uses, live in the glossary's own `@context` — as *\"a record outliving its dependency\"*, because `namespacesInUse()` reads **only** `processes/*.bpmn|dmn`. A vocabulary c", "file": "beans/defs/folio-assistant-2j09--the-undeclared-namespace-check-reads-only-bpmn-xml.md" }, { "id": "folio-assistant-2j2r", "title": "cmsl follow-up: ~44 readers of decl.directories miss entries declared from within", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-vke6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T17:24:45Z", "updatedAt": "2026-10-09T14:06:13Z", "preview": "## What\n\nSince cmsl step 1 (2026-09-30), `voices/` and `skills/lean/` are declared FROM WITHIN `skills/skills.json`, not in `.json`. `resolveDirectories` promotes them; but ~44 call sites read `readDeclaration(root).directories` directly and never see them.\n\nMeasured the same day: the UML overview deleted 5 voices diagrams, `check:wireframes` called 5 voices pages undeclared, and `viewer", "file": "beans/defs/folio-assistant-2j2r--cmsl-follow-up-44-readers-of-decldirectories-miss.md" }, { "id": "folio-assistant-2lzp", "title": "atomic-mass-gen-check.yml is qou folio residue in the platform repo — move it to litlfred/qou", "status": "completed", "type": "task", "priority": "low", "parent": "folio-assistant-vke6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T17:30:18Z", "updatedAt": "2026-10-07T17:18:00Z", "preview": "## Why\n`.github/workflows/atomic-mass-gen-check.yml` checks `content/quantum-observable-universe/lean/…`, which is not in this repository since the split. Its only live trigger here is an edit to the workflow file itself, and then it failed on a missing script (seen on PR #1284, which added its `# bpmn:` line). #1284 made the missing generator a stated \"not applicable\" notice rather than a red job", "file": "beans/defs/folio-assistant-2lzp--atomic-mass-gen-checkyml-is-qou-folio-residue-in-t.md" }, { "id": "folio-assistant-2ngl", "title": "MODE: developer — local server, CLI tools, one model for every workflow", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T08:55:36Z", "updatedAt": "2026-10-09T18:59:00Z", "preview": "From [#363](https://github.com/litlfred/folio-assistant/issues/363): \"developer\nmode: likely wants local server hostign the content, run CLI version of tools.\nuses single agent/model for all workflows\".\n\n## The strawperson position\n\nDeveloper mode is a **point in the topology product**, not a special code path:\n\n| axis | value |\n|---|---|\n| forge | local git only, or any — the mode does not care |", "file": "beans/defs/folio-assistant-2ngl--mode-developer-local-server-cli-tools-one-model-fo.md" }, { "id": "folio-assistant-2osx", "title": "Each instance publishes its own ns document for the block-kind classes it mints", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-7x5n", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T16:41:00Z", "updatedAt": "2026-10-09T16:27:00Z", "preview": "Follow-up to the vocabulary move (bean 0r7u, owner rulings 2026-10-06): block-kind classes are minted in the declaring instance's namespace (folio-assistant-core:Prose, folio-assistant-sci:Theorem, smart-base:Persona), but no instance publishes a document defining them, so the class IRIs do not dereference. The core ones never did (pre-existing gap); sci and smart-base are new namespaces with no n", "file": "beans/defs/folio-assistant-2osx--each-instance-publishes-its-own-ns-document-for-th.md" }, { "id": "folio-assistant-2q1p", "title": "RETIRED PREVIEWS COME BACK: the full-replace restore carries _retired as a record but never consults it, so a retirement is undone by the next main-site publish", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T12:00:53Z", "updatedAt": "2026-10-07T05:05:00Z", "preview": "Found 2026-10-02 by testing ONE preview retirement before dispatching 57 —\nwhich is the only reason it was found rather than paid for 57 times.\n\n## The failure, measured on `gh-pages`\n\n| commit | what it did |\n|---|---|\n| `a46c58ca9df` | `staging(cleanup): remove STAGING/claude-fervent-brahmagupta-rbwhzm` — **1615 files deleted** |\n| `93ffa64488c` | `docs(gh-pages): site from a07c3c66…` — **all 16", "file": "beans/defs/folio-assistant-2q1p--retired-previews-come-back-the-full-replace-restor.md" }, { "id": "folio-assistant-2r2n", "title": "Mobile: pages scroll sideways at phone width", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T10:36:13Z", "updatedAt": "2026-10-01T18:13:39Z", "preview": "A page wider than a 390 px viewport makes the reader pan sideways to reach columns or actions. Fix in the shared table/grid CSS: wrap, stack or collapse columns at narrow widths, and let only a table's own container scroll, with a visible cue.\n\nObserved on: `catalogue`, `folio`, `fsh-guts`, `kg-viewer`, `library`, `navbar`, `processes`, `skills-index`, `tools`, `translation-status`, `uploads` (see", "file": "beans/defs/folio-assistant-2r2n--mobile-pages-scroll-sideways-at-phone-width.md" }, { "id": "folio-assistant-2tfy", "title": "SEARCH COST: the lunr index is built on every page view — ~4.7 s CPU, ~315 MB heap before anyone searches", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-whlc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T09:04:02Z", "updatedAt": "2026-10-03T09:04:02Z", "preview": "Owner, 2026-10-03, choosing among measured options: **\"Both, lazy first\"** —\nPR 1 builds the search index on first focus of the search box; PR 2 shrinks\nthe index.\n\n## Measured (local `preview:site` build, headless Chromium, median of 3)\n\n| page | search index | main-thread script | task | JS heap |\n|---|---|---|---|---|\n| `/` | loaded on load (theme default) | 4,796 ms | 5,234 ms | 314 MB |\n| `/`", "file": "beans/defs/folio-assistant-2tfy--search-cost-the-lunr-index-is-built-on-every-page.md" }, { "id": "folio-assistant-2uj7", "title": "merge:guard follows Link next URLs in /repositories// form, which the agent proxy refuses (page 2 → 403)", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-nok9", "blocking": [], "blockedBy": [], "createdAt": "2026-10-05T05:11:59Z", "updatedAt": "2026-10-07T05:13:00Z", "preview": "Issue #2137. getAll in cat-harness/scripts/merge-guard.ts follows GitHub's Link rel=next, which names /repositories//...; the agent egress proxy refuses that form (403, 'Numeric-ID repository paths ... not supported'). Fix: normalise next to repos/{owner}/{repo}/.\n\n## Done when\n- [x] next-page URLs normalised; unreadable page still COULD NOT DETERMINE\n- [x] unit tests for page-2 normalised + p", "file": "beans/defs/folio-assistant-2uj7--mergeguard-follows-link-next-urls-in-repositoriesi.md" }, { "id": "folio-assistant-2upx", "title": "DOCUMENTATION: one SDO voice, RFC 2119 requirement levels, and the pages the knowledge graph is missing", "status": "in-progress", "type": "epic", "priority": "high", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-09-21T16:21:34Z", "updatedAt": "2026-10-04T15:12:15Z", "preview": "Owner, 2026-09-21. A large body of direction on documentation, captured here as the frame; the children carry the work.\n\n**The writing standard.** We are writing as an SDO writes — W3C, IHE. Glossary Terms are Capitalized throughout. Requirement levels follow RFC 2119 (MUST / MUST NOT / REQUIRED / SHALL / SHALL NOT / SHOULD / SHOULD NOT / RECOMMENDED / MAY / OPTIONAL). The RFC is queued at `upload", "file": "beans/defs/folio-assistant-2upx--documentation-one-sdo-voice-rfc-2119-requirement-l.md" }, { "id": "folio-assistant-2uui", "title": "WHO IRIS replica pages scroll sideways at desktop width", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T16:12:48Z", "updatedAt": "2026-09-30T17:17:27Z", "preview": "Found 2026-09-30 by the wireframe QA re-run on main 3779d5d27 (1280×800 and 1024).\n\n- who-iris/collection-collection-hq-publications.html: document 1366px wide at a 1280 viewport.\n- who-iris/community-list.html: 1503px at 1280.\n- A third page is too wide at 1024.\n\nCause: the download / 'Metadata record' cells do not wrap. #1592 (g9r2) made them wrap only below 640px, so phones are fixed and deskto", "file": "beans/defs/folio-assistant-2uui--who-iris-replica-pages-scroll-sideways-at-desktop.md" }, { "id": "folio-assistant-2vne", "title": "Folio handle back to top-centre as a short pill", "status": "completed", "type": "bug", "priority": "normal", "parent": "", "blocking": [], "blockedBy": [], "createdAt": "2026-09-27T05:19:36Z", "updatedAt": "2026-09-27T05:19:46Z", "preview": "Owner 2026-09-27: 'i want purple folio button, not on navbar but at top middle of display screen. not so tall'. Reverses 269z's navbar placement.\n\n## Summary of Changes\n\n- `docs-ui.js` `placeHandle`: the handle is appended to `body` again (fixed top centre) on every page, rail or theme sidebar alike.\n- `docs-ui.css`: pill 1.75rem tall (was 3.25rem), rounded bottom corners, 0.85rem label; `--in-nav", "file": "beans/defs/folio-assistant-2vne--folio-handle-back-to-top-centre-as-a-short-pill.md" }, { "id": "folio-assistant-2vpn", "title": "who-iris's top LHS navbar mark is the letter 'W', not an icon — and every harness's top mark must come from one mechanism", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-yg29", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T18:34:12Z", "updatedAt": "2026-10-05T04:57:30Z", "preview": "Owner, 2026-10-04, in session https://claude.ai/code/session_01Ga3HjmX3ag9vTgZWDSmsFi, on being asked to sign GOAL 3 off: *\"actually no, who-iris is missing top icon on LHS navbar. investigate why and if also try of other harnesses. all needs to be sconsisisten and consolidated\"* — with a screenshot of https://litlfred.github.io/folio-assistant/who-iris/ whose `.fa-nav-head` shows a slate tile wit", "file": "beans/defs/folio-assistant-2vpn--who-iriss-top-lhs-navbar-mark-is-the-letter-w-not.md" }, { "id": "folio-assistant-2yyh", "title": "SMART-BASE HARNESS: the WHO digital-health corpus, its methodologies and its voices", "status": "completed", "type": "epic", "priority": "high", "parent": "folio-assistant-vuip", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T08:35:00Z", "updatedAt": "2026-10-10T08:00:00Z", "preview": "Issue #877 https://github.com/litlfred/folio-assistant/issues/877\n\nThe owner, 2026-09-22: *\"WHO dgital health needs a once voice. also lots of process in DIIG for digital transformation. the DIIG is a basis for the digital trasnformation handbooks. ingest docs, make methodlolgies subgraphs for project management.digital trasnform processes\"*, and on placement: *\"same as other smart-* now in corpus", "file": "beans/defs/folio-assistant-2yyh--smart-base-harness-the-who-digital-health-corpus-i.md" }, { "id": "folio-assistant-3025", "title": "folio-paper-adapter: 44 of 47 skills have no role — the package is unmodelled, not untriaged", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-8jt6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T14:49:49Z", "updatedAt": "2026-10-09T21:56:00Z", "preview": "Measured 2026-09-20 while sampling for bean `y1w9`.\n\n## The finding\n\n`folio-paper-adapter` holds **47 skills and 44 are bound to nothing — 93 %.**\nThe 3 that are bound are scattered one-apiece across six unrelated roles\n(`reviewer` 2, `authoring-agent` 1, `onboarding-agent` 1, `ingestion-agent` 1,\n`evidence-agent` 1, `corpus` 1).\n\n**That pattern is the point.** Six roles holding one paper skill ea", "file": "beans/defs/folio-assistant-3025--folio-paper-adapter-44-of-47-skills-have-no-role-t.md" }, { "id": "folio-assistant-30hn", "title": "BPMN: 16 of 33 processes are strict by omission, not by decision", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T15:56:22Z", "updatedAt": "2026-10-09T16:03:00Z", "preview": "Found while auditing for `folio-assistant-haya`; recorded there in\n[SDLC process audit §4](../../cat-harness/docs/proposals/sdlc-process-audit.md).\n\n## The measurement\n\n`processes/` holds 33 `.bpmn` files. 17 carry a\n``; **16 do not.**\n\n grep -L 'folio:policy' processes/*.bpmn | wc -l # 16\n\n`src/workflow/process-model.ts:502` reads\n\n const enforcement = declar", "file": "beans/defs/folio-assistant-30hn--bpmn-16-of-33-processes-are-strict-by-omission-not.md" }, { "id": "folio-assistant-30jr", "title": "hfag's last three: declare the merge queue as a state graph, give it a viewer and tile, and record a train run", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-hfag", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T07:25:27Z", "updatedAt": "2026-10-21T17:26:00Z", "preview": "## Brief\n\nOwner, 2026-10-03: *\"do the three open hfag items\"*, then — when the tile turned out to be\nblocked on an integrity constraint rather than on effort — *\"1\"* (ship the other two, leave\nit open) *\"and 3\"* (surface the queue in the bean store rather than leaving it invisible).\n\n## What landed\n\n**The queue is a declared `state` graph.** Kind `merge-queue` in the registry — `holds: \"state\"`\nby", "file": "beans/defs/folio-assistant-30jr--hfags-last-three-declare-the-merge-queue-as-a-stat.md" }, { "id": "folio-assistant-319n", "title": "ZOD IS IN CAT-HARNESS: move bootstrap-tools' graph.ts and discussion.ts into cat-harness/schemas, retire bootstrap-tools if empty", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-88mg", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T22:52:17Z", "updatedAt": "2026-09-24T05:45:45Z", "preview": "Owner, 2026-09-23 (bean iwtn, ruling 4): *\"cat-harness/tools like renderer. zod is in cat-harness\"*. `model-registry.ts` already moved (PR #1200). The owner then chose to move the other two as well, in a follow-up PR after #1200 merges.\n\n## The rule (owner, 2026-09-23)\n\n> Zod - these are authoring tools for the KG, different tools can/will be used, but the core KG should be unchanged.\n\nThe Zod fil", "file": "beans/defs/folio-assistant-319n--zod-is-in-cat-harness-move-bootstrap-tools-graphts.md" }, { "id": "folio-assistant-31ni", "title": "Agent memory for every vendor: one memory source, one assembler, per-vendor agent files", "status": "in-progress", "type": "feature", "priority": "", "parent": "folio-assistant-7x5n", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T18:31:42Z", "updatedAt": "2026-10-23T20:40:00Z", "preview": "Owner question 2026-10-06: is there a Gemini/ChatGPT version of .claude/agent-memory to assemble too? Research (sources: vendor repos gemini-cli docs, openai/codex source, github/docs; Cursor/Antigravity UNVERIFIED): NO vendor but Claude has per-agent memory (memory: project -> .claude/agent-memory//MEMORY.md). Others have per-agent DEFINITION files whose body is the prompt: Gemini .gemini/", "file": "beans/defs/folio-assistant-31ni--agent-memory-for-every-vendor-one-memory-source-on.md" }, { "id": "folio-assistant-324x", "title": "regen reports docs:pages current while assets/beans/index.json is stale", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-hfag", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T07:18:41Z", "updatedAt": "2026-10-04T07:55:34Z", "preview": "Measured twice on #1955 (2026-10-04, after main merges that took main's copy of cat-harness/docs/assets/beans/index.json by the site-data pattern): the file lacked `edges` (bean vhqq), check:kind-validators:require-all failed in CI, yet `bun run cat regen` reported docs:pages:check current. Running `bun run cat docs:pages` rewrote it with 86 edges each time. So docs:pages:check does not compare th", "file": "beans/defs/folio-assistant-324x--regen-reports-docspages-current-while-assetsbeansi.md" }, { "id": "folio-assistant-32f6", "title": "Prefix the harness's special branches with cat- (qa-reports, lake-cache/*, state); gh-pages unchanged", "status": "in-progress", "type": "task", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T18:07:50Z", "updatedAt": "2026-10-22T14:20:00Z", "preview": "Owner, 2026-10-02, verbatim: \"need to prefix 'special' branches with cat-, cat-qa-reports, cat-fhir-ast, cat-lean-cacje (or whatever), not sure if any more. gh-pages stays as is\" — and later: \"(need to /coordinate and rename active branches)\".\n\nParent: fs43, not 7x5n — fs43 is the arc whose subject is the special branches (its P7, rva2, declares every special branch with one field); 7x5n is the se", "file": "beans/defs/folio-assistant-32f6--prefix-the-harnesss-special-branches-with-cat-qa-r.md" }, { "id": "folio-assistant-3432", "title": "BRANCH-ONLY RULING REQUEST: r0tm is a draft decision with a recommendation and a safe default, reachable from no ancestor of main", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T11:33:55Z", "updatedAt": "2026-10-09T13:25:00Z", "preview": "## The finding, and how it was measured\n\n`beans/defs/folio-assistant-r0tm--adapters-closure-*.md` exists on exactly one\nref — `origin/claude/yj6r-adapters-closure-plan` — and its commit\n(`974a0eac0ce`) is **not an ancestor of `origin/main`**:\n\n git merge-base --is-ancestor 974a0eac0ce origin/main -> non-zero\n git branch -a --contains 974a0eac0ce -> that one branch\n\nIt is ", "file": "beans/defs/folio-assistant-3432--branch-only-ruling-request-r0tm-is-a-draft-decisio.md" }, { "id": "folio-assistant-34cm", "title": "The committed site build under cat-harness/docs/ is 44% of all merge conflicts, and the publish workflow rebuilds it anyway", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-hfag", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T08:34:21Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "Owner, 2026-10-03: *\"I want you to examine each of the conflicted PR as they\ncome in to figure out how we can reduce the conflicts/ignore them.\"* This is\nthat measurement, taken over the whole open queue at once rather than per PR,\nbecause the answer is a distribution and not a property of any one PR.\n\n## The measurement\n\nEvery conflicted path of every open non-draft PR, against `main` `d089aca961", "file": "beans/defs/folio-assistant-34cm--the-committed-site-build-under-cat-harnessdocs-is.md" }, { "id": "folio-assistant-391j", "title": "MERGE SKEW: two green branches merged into a red main (prov-qaqc, 2026-09-24) — not a regen blind spot as first diagnosed; owner decides merge queue vs up-to-date vs fix-forward", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T06:06:37Z", "updatedAt": "2026-09-30T22:33:44Z", "preview": "Measured 2026-09-24. Main went red on 10b48aed (the #1246 merge) in `TypeScript — tests, lint, types`: the test `prov-qaqc: the real repository > the committed page and logs are current (what check:prov-qaqc gates)` failed. #1245 had added the report; 172b558d (#1190) then committed a new workflow instance without regenerating it. It was fixed by a pure regeneration in #1249.\n\n**Why `bun run cat r", "file": "beans/defs/folio-assistant-391j--regen-blind-spot-a-check-enforced-by-a-bun-test-no.md" }, { "id": "folio-assistant-398x", "title": "GUIDELINE DOMAIN: clinical / public health / health system strengthening, as a glossary scheme applied through classifications", "status": "in-progress", "type": "feature", "priority": "normal", "parent": "folio-assistant-0lmb", "blocking": [], "blockedBy": [], "createdAt": "2026-10-10T14:06:52Z", "updatedAt": "2026-10-23T16:42:00Z", "preview": "Owner, 2026-10-10: *\"these will all need to be a part of a SMART Guideline at some point and are one of clinical health, public health, or health system strengthening guidlines ... In any case these all will be be in the Glossary and can be sorted out later.\"*\n\n**Measured.** No code exists for this anywhere. The closest is smart-base `KGInterventionType` (DRAFT: `#public-health`, `#health-system`,", "file": "beans/defs/folio-assistant-398x--guideline-domain-clinical-public-health-health-sys.md" }, { "id": "folio-assistant-3clc", "title": "B10c (#1168): ModelId type + advisory check against the model registry", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T14:59:53Z", "updatedAt": "2026-09-30T15:20:06Z", "preview": "Owner 2026-09-30: Type + advisory. 532 free-string model ids across 5 fields; not-disclosed a declared value; bootstrap/models/models.json seeded with the 4 ids as validation: unverified.\n## Done when\n- [ ] ModelIdSchema on every model field\n- [ ] advisory check resolves ids against the registry\n\n\n\n## 2026-09-30 progress\n- [x] ModelIdSchema (bootstrap-tools/schemas/model-registry.ts) + NOT_DISCLOS", "file": "beans/defs/folio-assistant-3clc--b10c-1168-modelid-type-advisory-check-against-the.md" }, { "id": "folio-assistant-3dci", "title": "THEME: grumpy-cyborg-agents for cat-harness; plain grumpy-cat moves to folio-assistant-core", "status": "completed", "type": "feature", "priority": "", "parent": "", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T12:36:12Z", "updatedAt": "2026-09-24T12:36:12Z", "preview": "Owner 2026-09-24: supplied three crops (card 1254x1254, mobile 1024x1536, laptop 1659x948) of the grumpy cat in the sage hoodie on a sled pulled by four robot cats — \"import avatar theme\", \"use this for the cat-harness theme, the current plain grump cat them it is using should be for folio-assisnt-core\", named \"grumpy-cyborg-agents\".\n\n## Summary of Changes\n- Art at cat-harness/docs/assets/img/harn", "file": "beans/defs/folio-assistant-3dci--theme-grumpy-cyborg-agents-for-cat-harness-plain-g.md" }, { "id": "folio-assistant-3ds9", "title": "SPIKE: can CI and a fresh agent container both push to and read an orphan qa-reports branch through the proxy?", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T08:00:46Z", "updatedAt": "2026-10-01T08:53:39Z", "preview": "Arc `3fva`, proposal §4 item 1.2. **This spike can falsify part A**, so it goes before any mechanism is built.\n\nWrite path, as in lake-cache `cmd_seed`: hash-object, then mktree, then commit-tree, then push to a throwaway `qa-reports-spike`.\nRead path: `git fetch --depth=1 --filter=blob:none origin +qa-reports-spike:refs/qa-reports-read`, then `git show`.\n\nMeasure:\n- (1) push from a code-quality-g", "file": "beans/defs/folio-assistant-3ds9--spike-can-ci-and-a-fresh-agent-container-both-push.md" }, { "id": "folio-assistant-3f5f", "title": "KG CONTEXT: inSubgraph and partOf both map to dcterms:isPartOf — subgraph membership and structural containment collapse into one property", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-whlc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T09:46:16Z", "updatedAt": "2026-10-03T11:01:51Z", "preview": "Found by the c1m4 generator (gen-subgraph-jsonld.ts, 2026-10-03). In kg-export's context both `inSubgraph` (a node belongs to a named subgraph — a VIEW) and `partOf` (structural containment, e.g. an activity part of a process) map to `dcterms:isPartOf`. JSON-LD compaction therefore merges them into one property in the subgraph files, and a consumer cannot tell 'this skill is in subgraph skills/sdl", "file": "beans/defs/folio-assistant-3f5f--kg-context-insubgraph-and-partof-both-map-to-dcter.md" }, { "id": "folio-assistant-3fva", "title": "ARC: QA & test evidence off main — a commit-keyed qa-reports branch, and a test-plan → certification process", "status": "in-progress", "type": "epic", "priority": "high", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T07:59:35Z", "updatedAt": "2026-10-21T19:15:00Z", "preview": "Issue #1763. Proposal, workplan, dispatch map and decisions D1–D5: `cat-harness/docs/proposals/qa-reports-branch-and-test-process-2026-10-01.md`.\nSibling of the QA epic `1swy` (an epic cannot be parented under an epic here).\n\nOwner, 2026-10-01: do not pollute main with the QA subgraph; publish it on a\ndedicated branch, as caching is; tie QA into a strawperson test process (an agent\nor machine is t", "file": "beans/defs/folio-assistant-3fva--arc-qa-test-evidence-off-main-a-commit-keyed-qa-re.md" }, { "id": "folio-assistant-3hk4", "title": "QA-PUBLISH must publish real results once main holds none (5hox blocker)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T13:58:10Z", "updatedAt": "2026-10-21T19:15:00Z", "preview": "Found by the 5hox prep (2026-10-02, cat-harness/docs/proposals/5hox-removal-inventory.md). qa-publish publishes what the checkout holds under every declared qa directory. After 5hox removes the working copies, a fresh CI checkout holds almost nothing there (only the bootstrap kg-export sidecar), so every main/ entry would hold ~1 file and every --against main baseline would shrink to nothing.", "file": "beans/defs/folio-assistant-3hk4--qa-publish-must-publish-real-results-once-main-hol.md" }, { "id": "folio-assistant-3ka9", "title": "fhir-cache-seed-npm: follow transitive dependencies, and mirror smart-base's 4 missing packages", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T16:04:40Z", "updatedAt": "2026-10-08T01:15:00Z", "preview": "Split from 6mk7 on 2026-10-06 (session https://claude.ai/code/session_01EcBv3uwKYcnNbCC6BcPG92).\n\nSUSHI on smart-base still fails here, because 4 TRANSITIVE packages are not in the cache: crmi#2.0.0, cql#2.0.0, sdc#4.0.0 and terminology#7.3.0. The seeder reads only the IG's own `dependencies:`, so it never lists them. That means `--missing-out` never asks fhir-package-mirror for them either.\n\n## D", "file": "beans/defs/folio-assistant-3ka9--fhir-cache-seed-npm-follow-transitive-dependencies.md" }, { "id": "folio-assistant-3lbz", "title": "TOOLS: Zod schemas are not reachable as Tools, and kg-navigation has none — audit + analysis", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T05:16:23Z", "updatedAt": "2026-10-09T16:03:00Z", "preview": "**Asked by the owner, 2026-09-20**: *\"all zod usage is associated with skills\nand put in tools? (like validate node in graph, or so) in kg-navigation?\"* —\nand, on shape: *\"a Tool per Zod schema... no, but there should be common\npatterns (single pattern?) with some parameters more or less\"*.\n\nAnalysis: `cat-harness/docs/proposals/zod-schemas-as-tools.md`, attached to\n[#223](https://github.com/litlf", "file": "beans/defs/folio-assistant-3lbz--tools-zod-schemas-are-not-reachable-as-tools-and-k.md" }, { "id": "folio-assistant-3mo4", "title": "EXTRACTION: cleanMarkdownText leaves literal **** in a msgid when a code span is wrapped in emphasis, and eats the spacing when there are two", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T12:53:23Z", "updatedAt": "2026-10-07T05:07:00Z", "preview": "Found 2026-09-26 while measuring `lvk9`, and **pre-existing rather than introduced\nby it** — proven below.\n\n`cleanMarkdownText` strips an inline code span and strips emphasis markers, but\nwhen a code span is WRAPPED in emphasis it does neither cleanly, and it is not\nconsistent about which way it fails:\n\n| input | output |\n|---|---|\n| ``` **`clarity-defn-single`** — fail if x ``` | `\"**** — fail if", "file": "beans/defs/folio-assistant-3mo4--extraction-cleanmarkdowntext-leaves-literal-in-a-m.md" }, { "id": "folio-assistant-3o5b", "title": "TEST PROCESS: tester + certifier roles, system-under-test actor facet, test-plan-execution.bpmn, certification DMN, kg-audit criteria", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-3fva", "blocking": [ "folio-assistant-ff09" ], "blockedBy": [ "folio-assistant-ygzh" ], "createdAt": "2026-10-01T08:00:47Z", "updatedAt": "2026-10-02T05:51:43Z", "preview": "Arc `3fva`, proposal §3.2 and §4 items 5.4 and 5.5. Blocked on the TEST PLAN schema bean.\n\n`test-plan-execution.bpmn` has five lanes:\n- Requester / SUT: requests certification against plan P.\n- Tester: resolves the plan, binds the test data, executes, writes the run.\n- Untainted checker: re-executes a sample (`untainted-verification`).\n- Certifier: decides, using the DMN that `exitCriteria` names.", "file": "beans/defs/folio-assistant-3o5b--test-process-tester-certifier-roles-system-under-t.md" }, { "id": "folio-assistant-3oqj", "title": "FOUR KINDS NOTHING AUDITS: health, interaction, issue-marks, todos - 13 files, 0 criteria, 0 gates", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T05:44:17Z", "updatedAt": "2026-09-24T13:08:49Z", "preview": "Measured 2026-09-24 by `bun run cat audit:coverage` (bean `xutg`). Four declared\ngraph kinds hold files that **no `kg-audit` criterion and no gate declaring\n`@covers` reaches**:\n\n| kind | files | directories |\n|---|---|---|\n| `health` | 1 | `cat-harness/test/health/results` |\n| `interaction` | 1 | `interaction` |\n| `issue-marks` | 2 | `issue-marks` |\n| `todos` | 9 | `todos` |\n\n## This is an UPPER ", "file": "beans/defs/folio-assistant-3oqj--four-kinds-nothing-audits-health-interaction-issue.md" }, { "id": "folio-assistant-3ozg", "title": "bun test rewrites 72 committed script-sidecars, so bun run cat gates reports NOT clean on every branch, main included", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-27T05:06:01Z", "updatedAt": "2026-09-27T07:15:00Z", "preview": "Measured 2026-09-27 on pristine main `c6960465301`, and the same on 2026-09-26 at `8fb8a041a8e`.\n\n## What happens\n\nA full `bun test` rewrites all 72 files under `cat-harness/content/pipeline/script-sidecars/`, changing only three fields in each:\n\n last_run_at wall-clock time of this run\n last_run_sha the checked-out HEAD\n engine_version the local bun (bun-1.3.11 here; the committe", "file": "beans/defs/folio-assistant-3ozg--bun-test-rewrites-72-committed-script-sidecars-so.md" }, { "id": "folio-assistant-3p7c", "title": "MVP: bootstrap-provable layers — folio_init/instance-init against one layer alone, in an empty repo", "status": "completed", "type": "epic", "priority": "high", "parent": "folio-assistant-vuip", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T09:56:42Z", "updatedAt": "2026-10-04T09:56:42Z", "preview": "The MVP the owner ruled in `tndo` (2026-09-30), grouped so its progress can be read in one place.\n\n> **MVP = `folio_init` creates a working folio against that layer ALONE, in an empty repository.**\n\nRefined by the owner's 2026-10-01 ruling on `mer2`: for a layer with no adapter (`bootstrap`, `cat-harness`) the operation is an **instance-init**, not a folio — so \"working\" means the instance-init su", "file": "beans/defs/folio-assistant-3p7c--mvp-bootstrap-provable-layers-folio-initinstance-i.md" }, { "id": "folio-assistant-3q47", "title": "PROCESS WALK: deterministically set an agent's context from the KG at each task", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T17:04:01Z", "updatedAt": "2026-10-09T20:15:00Z", "preview": "**Uncaptured owner requirement**, same comment as `folio-assistant-cy4p` —\n[#363](https://github.com/litlfred/folio-assistant/issues/363#issuecomment-5740727385),\n09:20, not beaned until 17:0x.\n\n> use for skill to deterministcally walk agent through process flow,\n> detministitcally setting the memory/context using the KG\n> (subprocess/role overlays). worktiems are the beans that they carry with\n> ", "file": "beans/defs/folio-assistant-3q47--process-walk-deterministically-set-an-agents-conte.md" }, { "id": "folio-assistant-3r47", "title": "Drop the per-kind graph classes (KGraph, SkillGraph, VoiceGraph, …): a directory is a Subgraph holding GraphKind individuals", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-xsqm", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T10:04:38Z", "updatedAt": "2026-09-30T20:04:55Z", "preview": "## The ruling, owner 2026-09-30\n\nAsked *\"KGraph needs better name? CatHarness? … what is bootstrap's? consistent? why not just CatHarness? why need graph?\"*, then chose **\"Drop per-kind classes\"** over renaming.\n\n## Why (measured 2026-09-30)\n\n- bootstrap has no class per kind: a kind is a word (`skills`, `processes`) and one `bootstrap:GraphKind` individual, `bootstrap:graphKind/`; a declare", "file": "beans/defs/folio-assistant-3r47--drop-the-per-kind-graph-classes-kgraph-skillgraph.md" }, { "id": "folio-assistant-3sbm", "title": "TOOL RELEASES + PROFILES: named, versioned, hashed tool sets with per-tool runtimes; PROV-O log of what each run used; SPDX 3 as export", "status": "in-progress", "type": "feature", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-10-07T22:13:50Z", "updatedAt": "2026-10-21T21:32:00Z", "preview": "Issue #2481. Owner 2026-10-07: 'need named/versioned releases + sha of software. it's not a gate per se, but a logging of what version of the Tool was used for audit purpose'; 'different tools may need different instances of java, so cannot pin java globally'. Owner chose PROV log + SPDX export. Closes G4 of the SPDX proposal (bean sd5v); answers D1's missing consumer (bean ffv7).\n\n## Done when\n- ", "file": "beans/defs/folio-assistant-3sbm--tool-releases-profiles-named-versioned-hashed-tool.md" }, { "id": "folio-assistant-3sm2", "title": "A gate held red by decision cannot ratchet its own subject — so the catalogue question is asked of the CHANGE (the 25→36 growth claim is RETRACTED: it was a mutant's output)", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T19:39:54Z", "updatedAt": "2026-09-27T05:06:39Z", "preview": "Owner chose the route on 2026-09-26: **make the drift gate green rather than move\nthe masked gates out from behind it** (asked with four options, against `cpss`'s\n108 masked checks). Measuring first turned that into a finding: the gate cannot be\nbrought green by finishing a fixed list, because the list is growing faster than it\nis being cleared, and nothing stops it.\n\n## The arithmetic, measured o", "file": "beans/defs/folio-assistant-3sm2--the-t8g3-campaign-closed-24-catalogue-gaps-today-a.md" }, { "id": "folio-assistant-3spu", "title": "INGEST: the Handbook's text layer keeps 742 soft hyphens (U+00AD) that split words in library sections", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-slw1", "blocking": [], "blockedBy": [], "createdAt": "2026-09-29T21:51:12Z", "updatedAt": "2026-09-29T23:21:35Z", "preview": "9789241548960-eng carries 742 U+00AD in 186 section files ('recommenda­tions', 'organi­zation'); the two style guides carry none; 21 sections in other libraries do too. The LSI tokenizer now joins them (tokenizer v2), but the section TEXT still carries them, so summaries, translation extraction and any lexical grep see fragments. Done when: the PDF rungs normalise U+00AD (join across the line brea", "file": "beans/defs/folio-assistant-3spu--ingest-the-handbooks-text-layer-keeps-742-soft-hyp.md" }, { "id": "folio-assistant-3srh", "title": "TEN GATES HAVE NOT DECLARED @covers, so every audit-coverage verdict is an upper bound", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T05:44:18Z", "updatedAt": "2026-09-24T13:08:48Z", "preview": "`bun run cat audit:coverage` (bean `xutg`) reports coverage per declared graph kind\nfrom `@covers` lines gates carry in their own module docblocks. 118 scripts\ndeclare; **10 gates have not said**, measured 2026-09-24:\n\n- `check:agents-claims`\n- `check:context-emission`\n- `check:image-roles`\n- `check:avatar-instances`\n- `check:asset-roles`\n- `check:retired-front-matter`\n- `check:wireframes`\n- `chec", "file": "beans/defs/folio-assistant-3srh--ten-gates-have-not-declared-covers-so-every-audit.md" }, { "id": "folio-assistant-3tza", "title": "IG / SMART Guideline creation: new repo or in-repo sub-KG, and the sub-KG lifecycle (stage, seed, re-point, cutover) extracted from the smart-* separation", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T07:13:44Z", "updatedAt": "2026-10-08T04:30:00Z", "preview": "Owner, 2026-10-06, verbatim: *\"add new skill/process: when asked to create a new FHIR IG or Smart Guideline, that can either be done in a new repo or within the folio of an existing harnessed repo. the agent should determine users intent and act accordingly. if user creates it in a folio/, then it can move it to another repo later (this is essentially the create a sub-KG process and skills we have", "file": "beans/defs/folio-assistant-3tza--ig-smart-guideline-creation-new-repo-or-in-repo-su.md" }, { "id": "folio-assistant-3vc1", "title": "A nested bun install makes root tsc report 12 errors that CI does not have — and it is the only way to fix a stale nested lockfile", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T17:25:05Z", "updatedAt": "2026-09-27T10:04:12Z", "preview": "## The trap, measured 2026-09-26\n\n`cat-harness/adapters/mcp-server/bun.lock` was stale against its own manifest, so\nfixing it means `bun install` **in that directory** — there is no other way to\nregenerate a lockfile. That install creates a 56 MB\n`cat-harness/adapters/mcp-server/node_modules`, and from that moment the ROOT\n`bunx tsc --noEmit` reports **12 errors across 3 files under\n`cat-harness/a", "file": "beans/defs/folio-assistant-3vc1--a-nested-bun-install-makes-root-tsc-report-12-erro.md" }, { "id": "folio-assistant-3x2o", "title": "Two generated skill pages name a source directory that does not exist — the oe98 banner bug recurs for remote stubs", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T10:44:49Z", "updatedAt": "2026-09-27T06:12:28Z", "preview": "Found 2026-09-26 while verifying `hloc`'s link rewriting — by checking that\nevery rewritten link points at something that EXISTS, which also swept the\nlinks the generator already composed.\n\n## The measurement\n\n`cat-harness/docs/reference/skill-instructions/fhir-client-operations.md` and\n`smart-launch.md` each open with:\n\n > Generated from [`cat-harness/skills/remote-stubs/.md`](\n > h", "file": "beans/defs/folio-assistant-3x2o--two-generated-skill-pages-name-a-source-directory.md" }, { "id": "folio-assistant-423d", "title": "REVIEW COMMENTS: anchored to block id as 9gyz Findings, surviving moves — and the write path a static page lacks", "status": "todo", "type": "task", "priority": "normal", "parent": "folio-assistant-q4jm", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T21:02:55Z", "updatedAt": "2026-10-21T17:10:00Z", "preview": "Owner: *\"and then manage review comments etc.\"*\n\n**The real obstacle, measured.** `docs-ui.js` states it outright: *\"A\npublished page cannot write back to the repo\"* (lines 2077 and 2592). The\nreview page is static on gh-pages, so \"manage comments\" needs a WRITE PATH,\nand choosing it is the owner's decision:\n\n| option | anchors to | cost |\n|---|---|---|\n| A. GitHub PR review comments | a line of a", "file": "beans/defs/folio-assistant-423d--review-comments-anchored-to-block-id-as-9gyz-findi.md" }, { "id": "folio-assistant-42u5", "title": "#1955 hand-off to the merge manager: kept merged with main by its session, by hand", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-whlc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T12:31:39Z", "updatedAt": "2026-10-07T20:30:00Z", "preview": "**For the merge manager — #1955 is kept merged with `main` by its owning session, by hand.**\n\nThe `merge-main` bot cannot land `main` into this branch:\n- it refuses two **authored** conflicts, which have no declared pattern, correctly so (`merge-conflict-patterns`):\n - `cat-harness/scripts/check-retired-front-matter.ts` — both sides added an import; both are kept (this branch's `checkoutRootFor`,", "file": "beans/defs/folio-assistant-42u5--1955-hand-off-to-the-merge-manager-kept-merged-wit.md" }, { "id": "folio-assistant-437w", "title": "kg:subscribe refuses a fork whose declaration is one level down (smart-base, smart-trust, smart-immunizations): accept an upstreamPath", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-fnx4", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T19:43:32Z", "updatedAt": "2026-10-07T20:30:00Z", "preview": "## Why\n\nThe owner ruled 2026-10-06 that smart-trust, smart-base and smart-immunizations leave folio-assistant the way who-iris did: each fork is authoritative and folio-assistant keeps a REMOTE subscription to it via `kg:subscribe`, pinned, no submodules. Today `bun run cat kg:subscribe` refuses all three with `not-a-substrate: the root carries no Knowledge Graph declaration`, because each fork ne", "file": "beans/defs/folio-assistant-437w--kgsubscribe-refuses-a-fork-whose-declaration-is-on.md" }, { "id": "folio-assistant-4475", "title": "Run build-ig-site in an IG's own repository (smart-trust first), once folio-assistant is split", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-vke6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T19:07:16Z", "updatedAt": "2026-10-16T10:10:00Z", "preview": "From bamf (#1670). The owner, 2026-09-30: an IG's site is built by the IG's own repository, by instantiating the smart-guideline harness there, and existing repos are NOT instantiated until folio-assistant is split completely (epic vke6).\n\n## What is ready\nfhir-harness's build-ig-site stages an IG source repository as one just-the-docs Jekyll site, with the Publisher's site.data.fhir. Verified on ", "file": "beans/defs/folio-assistant-4475--run-build-ig-site-in-an-igs-own-repository-smart-t.md" }, { "id": "folio-assistant-4682", "title": "Agent memory target: Gemini CLI (not implemented — owner 2026-10-06)", "status": "todo", "type": "task", "priority": "", "parent": "folio-assistant-31ni", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T18:35:15Z", "updatedAt": "2026-10-06T18:35:15Z", "preview": "Owner 2026-10-06: bean up, DO NOT implement yet. Target format: .gemini/agents/.md (front matter + prompt body; @file imports allowed, depth 5). Generated by the shared memory assembler into a marked region, with a --check gate and a per-target budget that drops lowest-priority nodes and reports them. Research: see parent bean 31ni.\n\n_2026-10-20_ — Owner re-asked (session https://claude.ai/", "file": "beans/defs/folio-assistant-4682--agent-memory-target-gemini-cli-not-implemented-own.md" }, { "id": "folio-assistant-46l0", "title": "The same commit fails DIFFERENTLY in three environments, so a green run cannot be read as a green tree", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T06:33:48Z", "updatedAt": "2026-10-22T16:34:57Z", "preview": "Measured 2026-09-26 on ONE commit, `3c2523a8071`, three ways:\n\n| where | `bun test` failures |\n|---|---|\n| GitHub Actions | **7** |\n| a clean `git worktree` of that exact sha, full suite | **8** |\n| the two disagreeing files, run alone | **0** |\n\nThe sets are not nested. CI had two the worktree did not (`this repository's own\ncorpus`, `the real corpus`); the worktree had three CI did not (`FOLIO_R", "file": "beans/defs/folio-assistant-46l0--the-same-commit-fails-differently-in-three-environ.md" }, { "id": "folio-assistant-46nf", "title": "B2 (#1168): voices and user stories point at their role; Role.voice and Role.useCases out", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T22:20:00Z", "updatedAt": "2026-09-23T22:49:20Z", "preview": "Part of s4sp (plan B). The 35 role voice sentences become one-rule voice profiles in folio-assistant-core (activeIn.roles to the cat-harness role); the 107 useCases become scenarios/stories.json UserStory nodes pointing at their role. Role.skills typed SkillNameSchema. kg-audit role-declares-voice and role-has-use-cases replaced by role-has-story and story-role-resolves; check:voices resolves acti", "file": "beans/defs/folio-assistant-46nf--b2-1168-voices-and-user-stories-point-at-their-rol.md" }, { "id": "folio-assistant-46ob", "title": "Missing viewers: 100 navbar rows read 'no viewer yet' across 47 kinds on the live site", "status": "todo", "type": "epic", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-10-15T16:53:28Z", "updatedAt": "2026-10-22T14:30:00Z", "preview": "## The ask, owner 2026-10-15 (verbatim)\n\n> bean up for missing viewers\n\nContext: the owner's screenshot of `/en/cat-harness/library/` showed \"no viewer yet\" on Processes and other declared sub-graphs.\n\n## Measured\n\nFrom the deployed `gh-pages` (folio-assistant `ac9cba167`, 2026-10-15T16:46Z), page `en/cat-harness/library/index.html`, parsing the rendered rail: **210 rows, 100 of them \"no viewer ye", "file": "beans/defs/folio-assistant-46ob--missing-viewers-100-navbar-rows-read-no-viewer-yet.md" }, { "id": "folio-assistant-46qw", "title": "Rename state + fhir-ast/* to cat- names with rename-special-branch.sh (handoff to local agy agent)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T21:25:05Z", "updatedAt": "2026-10-02T21:33:31Z", "preview": "## Brief\n\nYou are the EXECUTOR for one task handed to you by the merge-steward cloud session (https://claude.ai/code/session_01ToWZR4RgTRCWeSsgxsSQfT). That session cannot push or delete branches: its environment blocks it. You can, because you run on the owner's machine with their git credentials.\n\nWHERE (you were sent here by one line naming this bean, branch and repo)\n- Repo: litlfred/folio", "file": "beans/defs/folio-assistant-46qw--rename-state-fhir-ast-to-cat-names-with-rename-spe.md" }, { "id": "folio-assistant-48a6", "title": "IG FORK SITES DRIFT: smart-trust and smart-base GitHub Pages lack the current harness chrome that smart-immunizations has", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T06:17:40Z", "updatedAt": "2026-10-16T10:10:00Z", "preview": "Reported by the owner, 2026-10-06 (session https://claude.ai/code/session_012qoycyCSGidZqW245vXhze), verbatim:\n\n> https://litlfred.github.io/smart-trust/ and https://litlfred.github.io/smart-base dont have updated harness on ghpages like https://litlfred.github.io/smart-immunizations does\n\n## What is known\n- **Reference (good):** https://litlfred.github.io/smart-immunizations/ shows the current ha", "file": "beans/defs/folio-assistant-48a6--ig-fork-sites-drift-smart-trust-and-smart-base-git.md" }, { "id": "folio-assistant-4ak5", "title": "PER-HARNESS KG EXPORT: every instance publishes its own JSON-LD + schema (split cat-harness.jsonld); root index.jsonld meta-skeleton at depth 1", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-whlc", "blocking": [], "blockedBy": [ "folio-assistant-2ee5", "folio-assistant-e0hw" ], "createdAt": "2026-10-02T21:36:29Z", "updatedAt": "2026-10-07T17:37:00Z", "preview": "Owner, 2026-10-02, verbatim: \"bean: please make sure harnesses generate json(ld)+schema for their semi-static KGs, their subgraphs etc. there are some rules in the KG corpsu but scattered. right now, i dont see folio-assistnat.jsonld in gh-pagaes for example. i think all subsumted into cat-harness? these should be split out into their component sub-grpah/harnesses. see related beans. maybe we also", "file": "beans/defs/folio-assistant-4ak5--per-harness-kg-export-every-instance-publishes-its.md" }, { "id": "folio-assistant-4ccr", "title": "WIREFRAME FINDINGS: usability and accessibility defects the as-is wireframes observed (#1023)", "status": "completed", "type": "epic", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T10:36:13Z", "updatedAt": "2026-10-12T18:26:29Z", "preview": "The as-is wireframes merged in #1032 record, for every declared harness visualiser, what the page actually does at 1280×800 (web) and 390×844 (mobile). The `## Findings` section of each `cat-harness/docs/wireframes//intent.md` holds only what was observed on the rendered page or read off its generator; nothing is inferred.\n\nThis epic turns those findings into work (owner, 2026-09-23: *\"make ", "file": "beans/defs/folio-assistant-4ccr--wireframe-findings-usability-and-accessibility-def.md" }, { "id": "folio-assistant-4d22", "title": "A bean completed AFTER its PR merges is orphaned by the next re-branch — twice in one session", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T11:49:49Z", "updatedAt": "2026-09-22T11:49:49Z", "preview": "Measured 2026-09-22, twice, on this branch.\n\n## The shape\n\nThe sequence that loses it:\n\n1. PR merges.\n2. Agent marks the bean `completed` and commits it to the branch.\n3. Agent re-branches from the merged main: `git checkout -B origin/main`.\n4. **The completion commit is now orphaned** — it was pushed to the remote\n branch but never merged, and the local branch no longer contains it.\n\nT", "file": "beans/defs/folio-assistant-4d22--a-bean-completed-after-its-pr-merges-is-orphaned-b.md" }, { "id": "folio-assistant-4dbr", "title": "Forge portability: GitLab and sovereign-compute as additional Tool nodes, not a sixth repo", "status": "in-progress", "type": "task", "priority": "normal", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-18T18:46:11Z", "updatedAt": "2026-10-21T17:35:00Z", "preview": "Opened 2026-09-18 at the owner's direction: \"github ok for now, bean for future\ngitlab / sovereign-compute use\".\n\n## The decision this records\n\nThe five-repo split proposal (#223) considered a sixth repository,\n`agent-harness-github`, holding the four PR-choreography skills\n(`prepare-merge-auto`, `pickup`, `watch`, `coordinate`) so the harness could run\non another forge or none. **Rejected in favo", "file": "beans/defs/folio-assistant-4dbr--forge-portability-gitlab-and-sovereign-compute-as.md" }, { "id": "folio-assistant-4dx7", "title": "LaTeX in the ✍ suggest editor: math macros, references, environments, validated against the KG", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-14T16:30:00Z", "updatedAt": "2026-10-16T07:05:00Z", "preview": "Owner, 2026-10-14 (session https://claude.ai/code/session_0147GQWs8akum21vSh5XYURR): *\"How does it handle latex? Don't implement now but add a bean to folio assistant sci.\"*\n\n## Context\n\nThe ✍ suggest action (a WYSIWYG Markdown editor beside ✎ edit on every Markdown block of every harness site, Milkdown Crepe loaded on first use, the change submitted as a prefilled GitHub issue) lives in the share", "file": "beans/defs/folio-assistant-4dx7--latex-in-the-suggest-editor-math-macros-referenc.md" }, { "id": "folio-assistant-4dxv", "title": "goal-review axis 1: the session listing is ordered by CREATION, so a window filter must page until creation predates the window", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T12:20:12Z", "updatedAt": "2026-09-30T00:27:10Z", "preview": "Found 2026-09-24 in a 1-day goal review. The rule concerned is `goal-review` §\"The six axes\", axis 1 (\"ASK THE SESSION API FIRST … List the sessions\").\n\n**Measured:** `list_sessions(mine, limit 40)` returned 40 sessions, 22 of them updated in the window, and `has_more: true`. The in-window rows were interleaved with sessions last active on 07-03, because the list is ordered by creation, not activi", "file": "beans/defs/folio-assistant-4dxv--goal-review-axis-1-the-session-listing-is-ordered.md" }, { "id": "folio-assistant-4fv8", "title": "Placement PR4: whole roles, stories, actors and capabilities move up to their owners", "status": "todo", "type": "task", "priority": "", "parent": "folio-assistant-iirv", "blocking": [], "blockedBy": [ "folio-assistant-pzwb" ], "createdAt": "2026-10-01T06:58:01Z", "updatedAt": "2026-10-15T16:49:43Z", "preview": "Placement proposal §2 (owner rulings 2026-09-30 §6; process: review → resolve issues → staged PRs). Parent of PR0/PR1 is `9umr`; PR2–PR9 sit under the separation epic because D4 (owner, 2026-10-01, option 3) interleaves them with the split stages into one ordered sequence. Every PR: references move with the subject; generated trees regenerated, never hand-edited; `beans/**` not rewritten; kg-qa si", "file": "beans/defs/folio-assistant-4fv8--placement-pr4-whole-roles-stories-actors-and-capab.md" }, { "id": "folio-assistant-4iey", "title": "PER-PLAN exit-criteria DMN: test-plan-execution looks up the plan's own decision table", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T05:51:37Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "Follow-up from 3o5b (owner, 2026-10-02: bean, build later). test-plan/v1 lets a plan name its exitCriteria DMN, but GW_ExitCriteria in test-plan-execution.bpmn always evaluates the platform default (processes/decisions/test-certification.dmn).\n\n## Do first (interim guard, small)\n- [x] kg:audit criterion: a plan whose exitCriteria names a table other than the default is flagged (major), so no plan ", "file": "beans/defs/folio-assistant-4iey--per-plan-exit-criteria-dmn-test-plan-execution-loo.md" }, { "id": "folio-assistant-4j86", "title": "STAGING CONE (file level): a preview rebuilds only what a PR's changed files can reach — general rule, in the skills", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [ "folio-assistant-nama" ], "createdAt": "2026-10-04T13:49:44Z", "updatedAt": "2026-10-05T11:39:15Z", "preview": "Owner, 2026-10-04: *\"staging rebuild only what is dependency cone of changes (general rule. update skills)\"*. Asked how fine-grained, the owner chose FILE LEVEL (option 1 of 3).\n\n## Today (measured 2026-10-04, read-only map)\nfeature-staging.yml's cut is a PREFIX MATCH, not a cone:\n- compose-docs.ts `carriedInstances()` (bean ga8a) carries a composed instance only if a changed path starts with its ", "file": "beans/defs/folio-assistant-4j86--staging-cone-file-level-a-preview-rebuilds-only-wh.md" }, { "id": "folio-assistant-4kj4", "title": "AVATARS: per-kind avatar, in and out of trash, both schemes, with a QA axis for coverage", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T11:08:23Z", "updatedAt": "2026-10-09T17:38:00Z", "preview": "Owner, 2026-09-19:\n\n> each content type should have an avatar in and out of trash. dark and\n> light mode. see work on themes by sibling for stickies.\n\nand, when asked whether that meant folio content types or `fsh-guts` node\nkinds:\n\n> if more than one kind then it fades through the avatars in a loop. issues\n> with that? also make sure needed accesabily modes. QA sidescares if avatar\n> thems not fu", "file": "beans/defs/folio-assistant-4kj4--avatars-per-kind-avatar-in-and-out-of-trash-both-s.md" }, { "id": "folio-assistant-4kq7", "title": "Adopt github/spec-kit: spec-before-code gate, one spec template, specs as issue comments, change-size splitting", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-21T14:46:51Z", "updatedAt": "2026-10-09T21:40:00Z", "preview": "Issue: https://github.com/litlfred/folio-assistant/issues/730\n\nOwner request 2026-09-21: align folio-assistant with github/spec-kit; no development before a spec exists; all specs in one template; specs posted as GitHub issue comments; oversized changes split into multiple issues so human adjudication on the PR stays feasible; follow industry standards for the split.\n\n## Blocking decision — RESOLV", "file": "beans/defs/folio-assistant-4kq7--adopt-githubspec-kit-spec-before-code-gate-one-spe.md" }, { "id": "folio-assistant-4l4d", "title": "GEN-DOCS-PAGES reads QA badges through qa-store (5hox blocker)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-3fva", "blocking": [ "folio-assistant-5hox" ], "blockedBy": [], "createdAt": "2026-10-02T13:58:10Z", "updatedAt": "2026-10-02T17:56:38Z", "preview": "Found by the 5hox prep (2026-10-02). docs:pages:check and check:ci-invocations go red with test/results/ absent: committed docs pages embed QA badges, and without the corpus gen-docs-pages writes 'not available', so every page reads stale. Either read the corpus through qa-store (fetch miss = UNKNOWN, never a stale page), or move the badges out of the committed pages into the build.\n\n## Done when\n", "file": "beans/defs/folio-assistant-4l4d--gen-docs-pages-reads-qa-badges-through-qa-store-5h.md" }, { "id": "folio-assistant-4l5q", "title": "B7b (#1168): a skill declares the directories it governs (graph-kinds:, governs:); coverage.skill removed", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T19:50:21Z", "updatedAt": "2026-09-24T19:59:59Z", "preview": "Owner 2026-09-24: 'A' for coverage.skill, and 'skill names the directory' for a domain skill over a generic-kind directory. 24 kind claims added to skills' graph-kinds:; 5 directories whose kinds are all generic named in governs: as /. scripts/skill-governance.ts reads both; a kind claim reaches the skill's instance, its dependents, and the platform. check-subgraph-coverage derives t", "file": "beans/defs/folio-assistant-4l5q--b7b-1168-a-skill-declares-the-directories-it-gover.md" }, { "id": "folio-assistant-4nfs", "title": "ONTOLOGIST / TERMINOLOGIST: skills, methodologies, tools and processes for reviewing content and classifying/coding it", "status": "in-progress", "type": "feature", "priority": "high", "parent": "folio-assistant-0lmb", "blocking": [], "blockedBy": [], "createdAt": "2026-10-10T14:07:07Z", "updatedAt": "2026-10-23T21:13:00Z", "preview": "Owner, 2026-10-10: *\"(add bean: need to provide ontologist/termonologist skills+methdologies+tools+processes when trying to review content and classify/code it)\"*.\n\n## What exists (measured 2026-10-10)\n- **Skills.** core `skills/library/cataloguing/ontologist.md` (scan → ambiguities → formal glossary → mapping); `glossary-terms.md` (\"reference first, define second\", candidate/authored/could-not-ex", "file": "beans/defs/folio-assistant-4nfs--ontologist-terminologist-skills-methodologies-tool.md" }, { "id": "folio-assistant-4pla", "title": "JSON-LD VOICE, MECHANICAL HALF: publish:verify checks object properties are links and no document leans on a remote @base", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-whlc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T07:23:50Z", "updatedAt": "2026-10-03T07:24:00Z", "preview": "The linked-data voice (bean 9y9j) holds 7 rules, all judgementOnly: a reviewer applies them, nothing checks the emitted documents. Two are decidable by a JSON-LD processor over the emitted files and belong in publish:verify's verifier set, which already runs on the built site before deploy (docs-site.yml) and on every PR's staging build (feature-staging.yml).\n\nMeasured 2026-10-03 over the 7,059 co", "file": "beans/defs/folio-assistant-4pla--json-ld-voice-mechanical-half-publishverify-checks.md" }, { "id": "folio-assistant-4pm8", "title": "PAGEFIND: evaluate as the search engine for the large-datasets subgraph", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T09:21:09Z", "updatedAt": "2026-09-25T00:00:00Z", "preview": "Owner, 2026-09-20: '4 bean up. for now give requrements/constraings/issues/pros/cons/otpons here only for analysis/discussion./ not durable.'\n\nSo this bean is the TRACKING ITEM and deliberately not the analysis. The requirements, constraints, pros, cons and options were given in chat for discussion and are explicitly NOT to be written into the repository yet. Recording that instruction here matter", "file": "beans/defs/folio-assistant-4pm8--pagefind-evaluate-as-the-search-engine-for-the-lar.md" }, { "id": "folio-assistant-4rbc", "title": "CI cone: enable the f017 input-hash skip under CI, seeded from main's last green run", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-hfag", "blocking": [], "blockedBy": [], "createdAt": "2026-10-07T19:39:28Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "Issue #2456. Owner approved 2026-10-07 ~19:40Z (\"1y, 2y, 3y, 4y\"). Measurements in the issue body.\n\n_2026-10-07T19:39:38Z_ — Claimed by claude/ci-cone-f017 (session https://claude.ai/code/session_013WbQekVypi9A6YQbLDXMmJ).\n\n## 2026-10-07: the approach as specified skips nothing, so no CI plumbing was built\n\n**The falsifier hit before any code was written.** Under f017's existing rules, a pull requ", "file": "beans/defs/folio-assistant-4rbc--ci-cone-enable-the-f017-input-hash-skip-under-ci-s.md" }, { "id": "folio-assistant-4sim", "title": "DCAT: describe a RELEASE of the glossary / terminology as a published dataset — held until the term model lands", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T11:48:00Z", "updatedAt": "2026-09-23T22:13:21Z", "preview": "Owner, 2026-09-22, ruling on #596's ruling 2: **\"DCAT: hold as bean as render of whatever we land on.\"**\n\nSo this is deliberately NOT built, and the reason is in the ruling's own words: DCAT renders *whatever we land on*, so building it before the term model is settled would render a moving target.\n\n## What DCAT is and is not\n\n**Not a term model.** It describes a `dcat:Dataset` and its `dcat:Distr", "file": "beans/defs/folio-assistant-4sim--dcat-describe-a-release-of-the-glossary-terminolog.md" }, { "id": "folio-assistant-4tel", "title": "merge:main has no conflict pattern for translation-qa results, so it aborts", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-d33q", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T18:25:31Z", "updatedAt": "2026-10-07T11:50:33Z", "preview": "Measured 2026-10-04 ~18:15Z merging main into `claude/gracious-mendel-du6nn8` after #2105: `merge:main` ABORTED on five conflicts it classified `[no declared pattern]` — `cat-harness/test/results/translation-qa/docs/installation.{ar,es,fr,ru,zh}.translation-qa.json`. Both sides had only regenerated them (55+/55- each side). Resolved by hand by taking main's copy, then regen-after-merge proved the ", "file": "beans/defs/folio-assistant-4tel--mergemain-has-no-conflict-pattern-for-translation.md" }, { "id": "folio-assistant-4tnt", "title": "folio-test main lags the pipeline: the dogfood folio does not exercise what folio_init writes today", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-vuip", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T15:10:09Z", "updatedAt": "2026-10-16T10:10:00Z", "preview": "Recorded from the qou work-plan analysis, 2026-10-04 (session https://claude.ai/code/session_01NdDGeP1SyShmoUssLuRZ91). Not started: recorded so the gap has an owner. litlfred/folio-test was stood up by 58h0 as the folio_init dogfood; its main is behind the current template and pipeline, so it no longer proves what it was made to prove. Re-measure the delta before acting.\n\n\nRe-parented 2026-10-04 ", "file": "beans/defs/folio-assistant-4tnt--folio-test-main-lags-the-pipeline-the-dogfood-foli.md" }, { "id": "folio-assistant-4tts", "title": "IG VARIABLES via Liquid: lift generate_smart_liquid.py into site.data and render IG pages from Jekyll/Liquid templates in one pass", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T12:36:15Z", "updatedAt": "2026-10-06T19:30:00Z", "preview": "Owner, 2026-10-01: *\"make use of jekyll/liquid templates in justthedocs pipeline. bean should be there on variable substitution\"*.\n\nNo bean tracked this before. It was P0's stated **open half** in two skills:\n- `ig-publisher-reduction` §\"P0 is a re-point\";\n- `ig-render-jekyll` §\"Where the variables come from\".\n\n**Settled design (from the skills; not redesigned here):**\n- **Lift** WHO's `generate_s", "file": "beans/defs/folio-assistant-4tts--ig-variables-via-liquid-lift-generate-smart-liquid.md" }, { "id": "folio-assistant-4u5d", "title": "IG artefact index built at publish by fhir-harness from the AST: no IG instance commits fhir-artifact-index/", "status": "in-progress", "type": "feature", "priority": "normal", "parent": "folio-assistant-n3h9", "blocking": [], "blockedBy": [], "createdAt": "2026-10-23T14:30:00Z", "updatedAt": "2026-10-23T14:36:00Z", "preview": "## Parent\n\nfolio-assistant-n3h9\n\nRelated: folio-assistant-p7mb (build derived artefacts at publish).\n\nOwner, 2026-10-23: \"fhir-artifact-index/ → shouldn't that be a part of fhir-harness?\" and \"1 and implement\" (plan, then implement). Plan: litlfred/fhir-harness#124. Thread \"fhir-harness IG viewers\".\n\n## Why\n\nsmart-trust and smart-base commit `fhir-artifact-index/` (index, menu, releases; 1.3 to 1.", "file": "beans/defs/folio-assistant-4u5d--ig-artifact-index-built-at-publish-by-fhir-harness.md" }, { "id": "folio-assistant-4vg7", "title": "WORK PLAN: two bean files declare the same id folio-assistant-t3n8 — every id-keyed reader sees one and silently loses the other", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T12:05:24Z", "updatedAt": "2026-10-03T12:34:37Z", "preview": "Found 2026-10-03 by the q8ar SQLite slice builder (its manifest lists t3n8 under duplicateIds). Files: `folio-assistant-t3n8--harness-display-names-every-instance-declares-a-hu.md` (parent yj32) and `folio-assistant-t3n8--the-archive-rung-stages-but-can-never-promote-and.md` (parent ahvw). Any reader keyed by id (beans CLI, claim-bean, the dashboard index, roadmap) resolves one and drops the other", "file": "beans/defs/folio-assistant-4vg7--work-plan-two-bean-files-declare-the-same-id-folio.md" }, { "id": "folio-assistant-4wzf", "title": "RENAME: agent-harness -> cat-harness (computable agentic testing harness)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-vke6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-18T19:24:30Z", "updatedAt": "2026-09-30T10:23:23Z", "preview": "## The ask\n\nOwner, in chat 2026-09-18:\n\n> need to rename agent-harness to \"cat-harness\" for\n> \"computable-agentic-testing-harness\"\n\nQueued rather than done inline: it arrived mid-`g6yr`, and the standing\ninstruction is to queue new tasks rather than pivot.\n\n## Why it is bigger than a sed\n\nThe name is load-bearing in several distinct namespaces, and they do not all\nrename on the same schedule.\n\n**O", "file": "beans/defs/folio-assistant-4wzf--rename-agent-harness-cat-harness-computable-agenti.md" }, { "id": "folio-assistant-4y2i", "title": "TOPOLOGY: sovereign cloud — hosted forge, jurisdiction-defined URLs, L4-L5 data stores", "status": "todo", "type": "task", "priority": "normal", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T08:55:37Z", "updatedAt": "2026-09-19T09:01:53Z", "preview": "**Strawperson.** From [#363](https://github.com/litlfred/folio-assistant/issues/363):\n\"soverign cloud: hosted git server, staging and publcation URLs follows different\ndeployment pattern (jusidiction defined). in L4-L5 context may have data stores\nlike hapi. for (public health) portals, (health) digital wallets),etc. or health\nworker decision spport. bean up to may strwperson and create issue to r", "file": "beans/defs/folio-assistant-4y2i--topology-sovereign-cloud-hosted-forge-jurisdiction.md" }, { "id": "folio-assistant-4z5o", "title": "Generated GitHub edit/source URLs escape the repo with ../ for every cross-instance figure", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T09:08:59Z", "updatedAt": "2026-10-07T11:50:39Z", "preview": "Generated pages compose GitHub URLs by joining an instance-relative path that leaves the instance root, producing links with `/main/../` in them:\n\n https://github.com/litlfred/folio-assistant/edit/main/../folio-assistant-core/processes/library/l1-document-ingestion.bpmn\n https://github.com/litlfred/folio-assistant/blob/main/folio-assistant-core/... (the blob form is correct)\n\nThe `edit/` f", "file": "beans/defs/folio-assistant-4z5o--generated-github-edit-source-urls-escape-the-repo-.md" }, { "id": "folio-assistant-502c", "title": "EXPERIMENT: forced along a BPMN vs managing it — do outcomes change, at what compute?", "status": "todo", "type": "task", "priority": "", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T17:04:01Z", "updatedAt": "2026-09-19T17:04:01Z", "preview": "**Uncaptured owner requirement**, same comment as `folio-assistant-cy4p`\nand `folio-assistant-3q47` —\n[#363](https://github.com/litlfred/folio-assistant/issues/363#issuecomment-5740727385),\n09:20, not beaned until 17:0x.\n\n> comparision scenarios: give LLM BPMN and force them along it, givem LLM\n> BPMN and let them manage workflow. do behaviour outcomes change? how\n> much time/compute was needed?\n\n", "file": "beans/defs/folio-assistant-502c--experiment-forced-along-a-bpmn-vs-managing-it-do-o.md" }, { "id": "folio-assistant-52cz", "title": "A CONFLICTED PR CREATES NO pull_request RUN, and g62s's headUnjudged cannot see it — the one absent-run case that needs no glob evaluation", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T14:07:18Z", "updatedAt": "2026-09-30T18:25:48Z", "preview": "## The incident, measured\n\nPR #1589, 2026-09-30. `Repository gates (hard)` failed on `86e403cdfdd`. I\nfixed the cause (`readme:subgraphs`, two stale directory READMEs), pushed twice,\nand **no `Code-quality gates` run was ever created for either new head.** The\nPR went on showing a red from a superseded commit.\n\nThe cause was not the queue and not a path filter. The PR was\n**`mergeable_state: \"dirt", "file": "beans/defs/folio-assistant-52cz--a-conflicted-pr-creates-no-pull-request-run-and-g6.md" }, { "id": "folio-assistant-52dz", "title": "|-", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T07:44:35Z", "updatedAt": "2026-09-20T09:12:10Z", "preview": "Owner, 2026-09-20: *\"whats using them? give more info\"*.\n\n## Answer: NOTHING is using them\n\nMeasured 2026-09-20 — and the measurement **corrected this bean's own\nframing**, which called them \"folio workflows living in the platform\".\n\n### 1. Nothing triggers them\n\nAll eight are **`workflow_dispatch` only**. `publish.yml` also declares\n`workflow_call`, and **no workflow calls it**. They fire if, and", "file": "beans/defs/folio-assistant-52dz--folio-workflows-live-in-the-platform-repo-do-they.md" }, { "id": "folio-assistant-52rd", "title": "goal-review axis 6: the CI-health tool covers the last 100 runs, which can be shorter than the review window", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T12:20:12Z", "updatedAt": "2026-09-30T00:27:10Z", "preview": "Found 2026-09-24 in a 1-day goal review. The rule concerned is `goal-review` §\"The six axes\", axis 6 (\"Runs in the window by workflow and outcome\"), and rule 1 (\"Could not determine is never rendered as clean\").\n\n**Measured:** `bun run cat check:ci-health` reported \"100 recent run(s) spanning 13.4h\" for a 24h window, plus 12.9h of Pages deployments. The first ~10.6h of the window had no CI verdict", "file": "beans/defs/folio-assistant-52rd--goal-review-axis-6-the-ci-health-tool-covers-the-l.md" }, { "id": "folio-assistant-53rv", "title": "NETWORK-ACCESS AGENT: fetch what this environment cannot reach (FHIR packages for the mirror, WHO guideline PDFs from IRIS, published IG outputs) and commit them to GitHub", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-10-10T15:06:54Z", "updatedAt": "2026-10-23T18:10:00Z", "preview": "Owner, 2026-10-10: *\"make bean for agent with netwrok access.\"* The network policy of this session's cloud environment refuses the hosts below with 403. The owner says changing that policy is not theirs to do. So the work below waits for an agent whose environment CAN reach them, or a person on an ordinary machine. Every item ends in a commit to a GitHub repository, which this environment can read", "file": "beans/defs/folio-assistant-53rv--network-access-agent-fetch-what-this-environment-c.md" }, { "id": "folio-assistant-55ao", "title": "Add a first-class `recommendation` block kind for document folios", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-0lmb", "blocking": [], "blockedBy": [], "createdAt": "2026-08-28T15:04:51Z", "updatedAt": "2026-10-10T17:04:26Z", "preview": "Split out of `ehve`, which added the `document` content type and deliberately\nstopped short of this.\n\n## Why\n\nAn L1 health-policy guidance note exists to state **recommendations**. A\nrecommendation is the block readers cite, implementers trace to, and reviewers\nsign off individually — it wants a label, a stable identity and a place in the\n`uses[]` graph, which is everything a `theorem` has. It is ", "file": "beans/defs/folio-assistant-55ao--add-a-first-class-recommendation-block-kind-for-do.md" }, { "id": "folio-assistant-57mn", "title": "qa-sweep loads contributions from cat-harness/ and registers 0 contributed checkers", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-29T23:57:45Z", "updatedAt": "2026-09-30T00:13:13Z", "preview": "Found while building #1492 (PR #1515). Owner chose a separate small PR (2026-09-29).\n\n**Measured:** qa-sweep.ts called loadContributions(INSTANCE_ROOT), and INSTANCE_ROOT is cat-harness/, which declares no dependencies. So a real sweep registered 0 contributed checkers, and folio-assistant-sci's proof-compile-cost and proof-no-cost-regression never ran. qa-checker-discovery.test.ts passed because ", "file": "beans/defs/folio-assistant-57mn--qa-sweep-loads-contributions-from-cat-harness-and.md" }, { "id": "folio-assistant-58ro", "title": "Merge refused: #1977 owed CI not green on its head", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-7x5n", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T14:30:42Z", "updatedAt": "2026-10-15T16:49:44Z", "preview": "PR #1977, handed back by the Merge Manager on 2026-10-04 (BACKFILLED at 14:50Z: the hand-back was made as a PR comment and a queue entry only, without this bean. Owner: \"Backfill + enforce\").\n\n## Refused\n- `merge:steward` at 14:30Z: Rule_HeadNotGreen: conflict declared, own CI missing-required.\n- Queue entry (`beans/queue/litlfred--folio-assistant--1977.json`, status `active`): STATUS UNKNOWN on t", "file": "beans/defs/folio-assistant-58ro--merge-refused-1977-owed-ci-not-green-on-its-head.md" }, { "id": "folio-assistant-5a3l", "title": "DEPLOYMENT: topologies and operating modes are two axes, not one list of modes", "status": "in-progress", "type": "epic", "priority": "high", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T08:52:35Z", "updatedAt": "2026-10-04T15:12:14Z", "preview": "Opened 2026-09-19 from [issue #363](https://github.com/litlfred/folio-assistant/issues/363),\n\"self-sovereign compute + test harness deployments\", at the owner's direction:\n\"bean up to may strwperson and create issue to review these\".\n\n## The claim this epic rests on\n\n#363 lists roughly ten deployment and utilization scenarios — local git only,\nprivate GitHub, sovereign cloud, self-sovereign, devel", "file": "beans/defs/folio-assistant-5a3l--deployment-topologies-and-operating-modes-are-two.md" }, { "id": "folio-assistant-5akx", "title": "KeyedBySchema and DirectoryStorageSchema were two spellings of one enum — route parsed in one and threw in the other", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T08:35:58Z", "updatedAt": "2026-10-08T01:15:00Z", "preview": "## What was measured\n\n`main@12b916e9a5`, probe calling both schemas on one `storage` value:\n\n A. DirectoryStorageSchema accepts keyedBy route? true\n B. resolveSubgraphSource THREW: ZodError |\n Invalid option: expected one of \"commit\"|\"tip\"\n\n`DirectoryStorageSchema.keyedBy` (`schemas/cat-harness.ts`) gained `\"route\"` with bean `1j3q`.\n`KeyedBySchema` (`schemas/subgraph-source.ts`), whic", "file": "beans/defs/folio-assistant-5akx--keyedbyschema-and-directorystorageschema-were-two.md" }, { "id": "folio-assistant-5blc", "title": "Digital Transformation Handbook (DTH): a voice, an L1 document subtype on Reference Architecture + DIIG concepts, and a source of methodologies / processes / glossary", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-qvxh", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T18:16:53Z", "updatedAt": "2026-10-07T17:38:00Z", "preview": "Owner, 2026-10-01 (during stage D, #1767), attaching 9789240116191-eng.pdf (DTH for health product catalogue): 'use for one voice for \"Digital Transformation Handbook\" a specific type of a L1 document that uses the Reference Architecture and DIIG concepts', then 'also as source of methodologies / processes / glossary'. Same session: commit 27078551 on main uploaded four WHO DTHs to uploads/ (prima", "file": "beans/defs/folio-assistant-5blc--digital-transformation-handbook-dth-a-voice-an-l1.md" }, { "id": "folio-assistant-5ea6", "title": "PDF INLINE VIEWER: pinned pdf.js generic viewer on the site, embed on who-iris item pages, as skill + Tool node", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T18:40:22Z", "updatedAt": "2026-10-05T05:35:26Z", "preview": "Owner request 2026-10-04: lightweight inline PDF viewer (search, scroll, jump to page, print, download) for CDN-hosted PDFs, e.g. who-iris item pages. Owner chose option 2 (pdf.js generic viewer copied onto the site) and asked for it as a skill and a tool.\n\n## Todo\n- [x] vendor script: pinned pdf.js release, sha256-verified, pruned, installed into _site at build time (not committed)\n- [x] cross-or", "file": "beans/defs/folio-assistant-5ea6--pdf-inline-viewer-pinned-pdfjs-generic-viewer-on-t.md" }, { "id": "folio-assistant-5ge1", "title": "MERGE GATE IS SPECIFIED AS BLOCKING, OWNER RULED WARN-ONLY: reconcile merge-gate-2026-10-02.md and nok9 with the 2026-10-02 ruling", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-0ipy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T23:32:19Z", "updatedAt": "2026-10-03T00:02:20Z", "preview": "## The conflict, both sides measured 2026-10-02\n\n**The owner ruled WARN:** asked how to handle agentic code review before merge,\nthe owner answered — *\"do ingesion and analsysi. then propose. dont want hard\ngate (at least not for now, lots of backlog on content nodes) but do want\nwarn.\"* Two reasons given, and the second is the binding one: a hard gate over\na content backlog blocks work that was a", "file": "beans/defs/folio-assistant-5ge1--merge-gate-is-specified-as-blocking-owner-ruled-wa.md" }, { "id": "folio-assistant-5gqn", "title": "MERGE FRICTION: beans index.json should join oxka -merge set; harness.json must NOT", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-27T06:16:21Z", "updatedAt": "2026-10-07T04:37:00Z", "preview": "Measured 2026-09-27 while driving #1439: **three** merges of `origin/main` in\nabout forty minutes, and `docs/assets/beans/index.json` conflicted on the third\none. `oxka` fixed this for three files and marked them\n`linguist-generated=true -diff -merge` in `.gitattributes`. This file is a fourth\nwith the same shape and is not marked.\n\n| file | conflicts paid today | marked `-merge`? |\n|---|---|---|\n", "file": "beans/defs/folio-assistant-5gqn--merge-friction-beans-indexjson-should-join-oxka-me.md" }, { "id": "folio-assistant-5hox", "title": "REMOVE moved QA files from main — only on the owner's explicit go, after the branch holds a hash-verified copy", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [ "folio-assistant-129h", "folio-assistant-4l4d", "folio-assistant-f3bh", "folio-assistant-lb8h" ], "createdAt": "2026-10-01T08:00:47Z", "updatedAt": "2026-10-16T10:30:00Z", "preview": "Arc `3fva`, proposal §4 item 3.7 and decision D4. **Never on an agent's own initiative** (`deletion-requires-confirmation`).\n\nPreconditions:\n- every reader is migrated;\n- the gates have been green on the branch for 7 days (D4 default);\n- the `qa-reports:main/` tree is byte-identical to main's `test/results` for every moved path, checked by blob hash;\n- attestations stay in place (D2 default)", "file": "beans/defs/folio-assistant-5hox--remove-moved-qa-files-from-main-only-on-the-owners.md" }, { "id": "folio-assistant-5ki9", "title": "Merge refused: #2082 owed CI not green on its head", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-qvxh", "blocking": [ "folio-assistant-70zt" ], "blockedBy": [], "createdAt": "2026-10-04T14:42:21Z", "updatedAt": "2026-10-07T14:55:47Z", "preview": "PR #2082, handed back by the Merge Manager on 2026-10-04 (BACKFILLED at 14:50Z: the hand-back was made as a PR comment and a queue entry only, without this bean. Owner: \"Backfill + enforce\").\n\n## Refused\n- `merge:steward` at 14:30Z: Rule_HeadNotGreen: conflict clean, own CI red.\n- Queue entry (`beans/queue/litlfred--folio-assistant--2082.json`, status `active`): Owner-approved 2026-10-04 ~14:00Z (", "file": "beans/defs/folio-assistant-5ki9--merge-refused-2082-owed-ci-not-green-on-its-head.md" }, { "id": "folio-assistant-5o5c", "title": "Root-dependent platform outputs judged on a foreign content root (translation:obsolete, state:visualizer, processes:render)", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-21T15:37:07Z", "updatedAt": "2026-10-23T09:00:00Z", "preview": "Measured on litlfred/smart-immunizations#66 (run 39619275196, every layer at main, 2026-10-21). Three gates fail on a content root for platform artefacts that only the DECLARING checkout produces, so no data PR can make them pass everywhere — committing what smart-immunizations' root builds makes folio-assistant's own run stale, and the reverse.\n\nSame family as cat-tools-x5e0 (a content index's ga", "file": "beans/defs/folio-assistant-5o5c--root-dependent-platform-outputs-judged-on-a-foreig.md" }, { "id": "folio-assistant-5p4m", "title": "CI: remove the two 1-s roll-up jobs (typescript, e2e) — 7fu5 leftover, owner swaps required checks", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-hfag", "blocking": [], "blockedBy": [], "createdAt": "2026-10-07T19:39:28Z", "updatedAt": "2026-10-08T01:15:00Z", "preview": "Issue #2456. Owner approved 2026-10-07 ~19:40Z (\"1y, 2y, 3y, 4y\"). Measurements in the issue body.\n\n_2026-10-07T19:39:38Z_ — Claimed by claude/ci-runner-budget (session https://claude.ai/code/session_013WbQekVypi9A6YQbLDXMmJ).\n\n## Done on claude/ci-runner-budget (PR #2457), 2026-10-07\n\n- Removed jobs `typescript` and `e2e` (commit 35a8229de93). Nothing outside the workflow read the names except th", "file": "beans/defs/folio-assistant-5p4m--ci-remove-the-two-1-s-roll-up-jobs-typescript-e2e.md" }, { "id": "folio-assistant-5qq3", "title": "REGEN SCOPE: the browser jobs' artefacts are outside regen's fast set, and a stale one only surfaces in CI", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-2upx", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T18:53:17Z", "updatedAt": "2026-10-03T18:39:06Z", "preview": "Found 2026-09-30 on PR #1633, by CI rather than by `regen`.\n\n`End-to-end + accessibility (hard)` failed on `render:bpmn:check` with three\nstale drawings — `bootstrap/processes/{discussion,initialize-harness,log-message}.svg`\n— on a tree where `bun run cat regen` had just reported a fixed point.\n\n## This is NOT the same defect as `check:X` (the prefix-convention false clean)\n\nWorth separating, beca", "file": "beans/defs/folio-assistant-5qq3--regen-scope-the-browser-jobs-artefacts-are-outside.md" }, { "id": "folio-assistant-5qy8", "title": "QA SIDECAR LOCATION CONTRADICTION: the folio_init template commits *.qa.json while AGENTS.md puts QA on the qa-reports branch", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T15:10:09Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "Recorded from the qou work-plan analysis, 2026-10-04 (session https://claude.ai/code/session_01NdDGeP1SyShmoUssLuRZ91). Not started: recorded so the gap has an owner. The owner ruled 2026-10-04 for qou: use the cat-harness qa-reports branch. The folio_init template (cat-harness/templates/) still writes a layout that commits sidecars on main. Reconcile the template with skills/sdlc/sdlc-core/qa-rep", "file": "beans/defs/folio-assistant-5qy8--qa-sidecar-location-contradiction-the-folio-init-t.md" }, { "id": "folio-assistant-5rmf", "title": "NAVBAR QR ICON GONE: the LHS top row's QR code for the current page no longer appears, though its generator still loads", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-9rq1", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T06:51:27Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "Owner, 2026-10-06 (https://claude.ai/code/session_012qoycyCSGidZqW245vXhze), verbatim:\n\n> LHS top navbar used to have a QR code icon that would make a QR code for the currently viewed page and show it inside the LHS navbar.\n\n## Measured\n- The generator is still shipped. `cat-harness/docs/_includes/head_custom.html:181` documents *\"Site UI: a QR of the current page in the sidebar header\"*, and line", "file": "beans/defs/folio-assistant-5rmf--navbar-qr-icon-gone-the-lhs-top-rows-qr-code-for-t.md" }, { "id": "folio-assistant-5ulq", "title": "STRICT IN NAME ONLY: 9 workflow instances against 755 merges, and no gate reads beans/workflows at all", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-0ipy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T23:49:21Z", "updatedAt": "2026-10-12T09:36:00Z", "preview": "## Measured 2026-10-02 at `871b1da1c`\n\n| | |\n|---|---|\n| instance files in `beans/workflows/` | **9** |\n| first-parent merge commits on `origin/main` since 2026-09-20T18:00Z | **755** |\n| gates or workflows referencing `beans/workflows` | **0** (grep over `.github/workflows/` and `cat-harness/scripts/check-*.ts`) |\n| bean `vlhk`, which made the rule mandatory | `status: completed`, archived |\n\n`AG", "file": "beans/defs/folio-assistant-5ulq--strict-in-name-only-9-workflow-instances-against-7.md" }, { "id": "folio-assistant-5uuf", "title": "FOLIO PREVIEW BEFORE-SIDE: no main-site publish, so before pictures and 'view on main' are empty; stacked PRs compare with the wrong base", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-q4jm", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T18:06:56Z", "updatedAt": "2026-10-07T19:30:00Z", "preview": "Found by ojcx's real run, 2026-09-23 (litlfred/folio-test#6). The visual diff's \"before\" picture, and the review page's \"view on main\" link, both read the published main site at the root of `gh-pages`. Two gaps follow:\n- **A folio has no main-site publish.** Neither `init-folio` nor `folio-staging.yml` publishes `main`'s document site to the `gh-pages` root. So every before picture reads *\"Its pag", "file": "beans/defs/folio-assistant-5uuf--folio-preview-before-side-no-main-site-publish-so.md" }, { "id": "folio-assistant-5uyl", "title": "Extract the immunizations L1 graph from DAK Component 1 into smart-immunizations/library", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ioa4", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T08:58:14Z", "updatedAt": "2026-10-21T18:48:00Z", "preview": "## Done when\n- [x] the DAK PDF is ingested by folio-assistant and its entry placed under smart-immunizations/library/\n- [x] a generic extractor reads Component 1 (§1.1 interventions, §1.2 cited guidance) and resolves each (n) against the reference list\n- [x] the graph validates with smart-kg tools/validate.mjs at WHO main AND the Zod validator in smart-base\n- [x] every cited reference with a retri", "file": "beans/defs/folio-assistant-5uyl--extract-the-immunizations-l1-graph-from-dak-compon.md" }, { "id": "folio-assistant-5vo9", "title": "ADJUDICATION AS A TYPED SUBPROCESS: one judgement contract, reusable in every BPMN", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T06:15:42Z", "updatedAt": "2026-10-09T15:38:00Z", "preview": "Formalize 'use judgement' as a declared subprocess with a typed I/O contract, so a non-deterministic step is marked as one rather than left implicit. Owner's spec 2026-09-23. A CONCRETE INSTANCE ALREADY EXISTS (processes/adjudication.bpmn) — this generalises it rather than starting over.\n\n## The owner's specification, 2026-09-23, verbatim in substance\n\n> need formalized adjudication process so the", "file": "beans/defs/folio-assistant-5vo9--adjudication-as-a-typed-subprocess-one-judgement-c.md" }, { "id": "folio-assistant-5xfr", "title": "qou's pin bump must rename folio.config.json AND add qaAxes in the same commit, or it silently loses both", "status": "todo", "type": "task", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T14:37:58Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "## What is true TODAY, measured, because I got it wrong first\n\nI reported in session chat that \"qou is running without that axis as of this\nmerge\", after fencing `detangler-archimedean-wall` behind `qaAxes`. **That was\nwrong, and the reason is the pin.**\n\n`qou/.folio-assistant-pin` → `df28d02` (2026-09-07). Read at that commit:\n\n- `folioOptionalAxes()` exists and reads `join(findContentRepoRoot(),", "file": "beans/defs/folio-assistant-5xfr--qous-pin-bump-must-rename-folioconfigjson-and-add.md" }, { "id": "folio-assistant-5xzc", "title": "QA: block ids in the folio/ graph are unique and stable across render, move and re-ingest — the precondition every review view keys on", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-q4jm", "blocking": [ "folio-assistant-jwox", "folio-assistant-v26p" ], "blockedBy": [], "createdAt": "2026-09-22T21:02:54Z", "updatedAt": "2026-09-29T18:14:49Z", "preview": "Owner: *\"blocks are in the knoledge graph folio/. QA, to manage IDs\"*.\n\n**Why first.** Every other child keys on the block id. A diff matches blocks\nacross `main` and STAGING by id, a review comment anchors to an id, and a heat\nmap aggregates by id. If ids drift, the diff reads everything as removed and\nre-added, and every comment orphans. That is how this whole epic fails.\n\n**Measured 2026-09-22.", "file": "beans/defs/folio-assistant-5xzc--qa-block-ids-in-the-folio-graph-are-unique-and-sta.md" }, { "id": "folio-assistant-5yhm", "title": "TERMINOLOGY: identification, mapping and adjudication — a glossary term is CHECKED against an existing terminology, never only minted", "status": "in-progress", "type": "feature", "priority": "normal", "parent": "folio-assistant-0lmb", "blocking": [], "blockedBy": [], "createdAt": "2026-09-25T04:51:06Z", "updatedAt": "2026-10-02T08:01:23Z", "preview": "Owner, 2026-09-25: *\"extracting exsiting glossary needs compaision to existing\ntermonology/coding. there needs to be a whole set of skills on terminiology\nindetificatio. mapping. ajdudication. bean up as todo. we will integrate, for\nexample, OCL at some point as a Tool\"*.\n\n## RE-SCOPED 2026-09-29 against what has since landed on `main`\n\nThis bean was written when the glossary only extracted. Three", "file": "beans/defs/folio-assistant-5yhm--terminology-identification-mapping-and-adjudicatio.md" }, { "id": "folio-assistant-5zs1", "title": "Agent memory target: OpenAI Codex CLI (not implemented — owner 2026-10-06)", "status": "todo", "type": "task", "priority": "", "parent": "folio-assistant-31ni", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T18:35:16Z", "updatedAt": "2026-10-06T18:35:16Z", "preview": "Owner 2026-10-06: bean up, DO NOT implement yet. Target format: .codex/agents/.toml developer_instructions; repo-wide AGENTS.md (32 KiB summed chain). Generated by the shared memory assembler into a marked region, with a --check gate and a per-target budget that drops lowest-priority nodes and reports them. Research: see parent bean 31ni.\n\n_2026-10-20_ — Owner re-asked (session https://clau", "file": "beans/defs/folio-assistant-5zs1--agent-memory-target-openai-codex-cli-not-implement.md" }, { "id": "folio-assistant-603s", "title": "LANDING: the LHS navbar is one themed section per instance, scanned from the root, rendered in dependency order", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-yj32", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T12:32:20Z", "updatedAt": "2026-10-04T19:55:08Z", "preview": "Owner, 2026-09-20, verbatim — the structure is specific enough that paraphrase\nwould lose the parts that constrain the design:\n\n> then landing page rendering is now cleaer... scan the root of the repo, look\n> for .config.json as named instances. do their rendering (in\n> depdendcy ordering). on LHS navbar, there is global section (w/ themed color)\n> for each harness. when collapsed on LHS ", "file": "beans/defs/folio-assistant-603s--landing-the-lhs-navbar-is-one-themed-section-per-i.md" }, { "id": "folio-assistant-61ca", "title": "B5-fix (#1168): a GitHub workflow names the BPMN it implements; 9 and flipped", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T17:22:12Z", "updatedAt": "2026-09-24T17:47:57Z", "preview": "## Why\n9 `` are arrow-direction findings: the diagram names its implementation. Owner decision 2026-09-24: \"Workflow names its BPMN\".\n\n## Plan\n- Each `.github/workflows/*.yml` carries `# bpmn: cat-harness/processes/.bpmn`; each job implementing a node carries `# bpmn-node: `.\n- `check-workflow-coverage.ts`, `prose-code-pairs.ts`, `pair-claim", "file": "beans/defs/folio-assistant-61ca--b5-fix-1168-a-github-workflow-names-the-bpmn-it-im.md" }, { "id": "folio-assistant-61n5", "title": "qa-results parity test fails intermittently under bun run cat gates (2 of 4 local runs), never alone or in CI", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T07:09:46Z", "updatedAt": "2026-09-24T12:40:45Z", "preview": "The test \"every family in the committed result matches the export's own field\" in cat-harness/scripts/tests/qa-results.test.ts failed in 2 of 4 local `bun run cat gates` runs on 2026-09-23, on diffs that don't touch kg-export (the a1lq and 423d work). It passed 3 of 3 plain `bun test` runs, passes when run alone, and passed in CI every time.\n\nIt compares the committed kg-export.qa-results.json aga", "file": "beans/defs/folio-assistant-61n5--qa-results-parity-test-fails-intermittently-under.md" }, { "id": "folio-assistant-61t6", "title": "fsh-guts checks treat a separated subtree as one frozen item described by its note", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-3tza", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T20:21:29Z", "updatedAt": "2026-10-08T04:30:00Z", "preview": "cf210f4 deposited fsh-guts/separated/{smart-base,smart-trust,smart-immunizations}/ (~7.7k files) on cat/cat-harness/fsh-guts and every PR's CI went red (export @context lacked `data`, render-pipeline tests and the page walked every file, materialized-fixity judged a frozen copy's index.json, bean refs to hupw). 611a3fd held the cutover back until #2320 merges. Platform must treat separated//", "file": "beans/defs/folio-assistant-61t6--fsh-guts-checks-treat-a-separated-subtree-as-one-f.md" }, { "id": "folio-assistant-61tg", "title": "TOPOLOGY: self-sovereign — own infrastructure, inward-facing, wallets and DAK intake", "status": "todo", "type": "task", "priority": "normal", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T08:55:37Z", "updatedAt": "2026-09-19T09:01:53Z", "preview": "**Strawperson.** From [#363](https://github.com/litlfred/folio-assistant/issues/363):\n\"self-soveringn mode: like soveregion cloud but hosted on own infrastructure, in\nL4-L5 context may have data stores like hapi. not designed for external facing\nservices. more for managing wallets. have incoming DAK content. bean up to may\nstrwperson and create issue to review these.\"\n\n## Strawperson position, to ", "file": "beans/defs/folio-assistant-61tg--topology-self-sovereign-own-infrastructure-inward.md" }, { "id": "folio-assistant-6306", "title": "METHODOLOGY NODES for the three W3C sources: PROV-O (QA provenance), ODRL (permissions), JSON-LD (KG serialisation)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-scfh", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T08:12:46Z", "updatedAt": "2026-10-01T09:20:34Z", "preview": "#1744 merged red: library-ref.test.ts 'the PLATFORM's library holds ONLY sources its methodologies cite' fails on w3c-2013-prov-o, w3c-2018-odrl-model-2-2, w3c-2020-json-ld-1-1. Owner chose option C (#1614): write three real methodology nodes in cat-harness/methodologies/, each grounded in the held source and in how the platform actually uses the standard.\n\n## Done when\n- [x] prov-o node, evidence", "file": "beans/defs/folio-assistant-6306--methodology-nodes-for-the-three-w3c-sources-prov-o.md" }, { "id": "folio-assistant-63es", "title": "DEPLOY-ONLY CHECKS: a built-site check that runs only on main lets a PR stay green and break every publish", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T06:31:51Z", "updatedAt": "2026-10-01T07:06:44Z", "preview": "Follow-on to #1726 / #1730. Owner 2026-10-01: build the early catch.\n\ncheck:escaped-markup and check:maintained-artefacts ran only in docs-site.yml (the main deploy). A code span wrapped so a line began '' (introduced in #1615) broke every publish for ~6h while each PR's 'stage' stayed green.\n\n## Measured\n- Ran both checks on real staging trees from gh-pages: pass on this branch, #1615 and ", "file": "beans/defs/folio-assistant-63es--deploy-only-checks-a-built-site-check-that-runs-on.md" }, { "id": "folio-assistant-63q4", "title": "Bibliography: Reference KG nodes, a path-limited bibliography viewer (like the glossary), exports in BibTeX / CSL-JSON / RIS / Dublin Core, per asset and per subgraph", "status": "in-progress", "type": "feature", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-10-14T17:45:00Z", "updatedAt": "2026-10-23T20:44:00Z", "preview": "## The ask, owner 2026-10-14 (verbatim)\n\n> what is the general bibliography support? look at who-iris - its dublin core recores are a basis for a bibliography - there should be abibliogrpahy viewere that allows view/export in multiple formats of bib items in the folio. those citations be from library/ contents or extracted from a paper reference in the library or fiolo/. you should also be able to", "file": "beans/defs/folio-assistant-63q4--bibliography-reference-nodes-viewer-and-exports.md" }, { "id": "folio-assistant-63wl", "title": "Placement PR3: BPMN and DMN move up to their owners; harness processes regroup by concern", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-iirv", "blocking": [ "folio-assistant-apcg" ], "blockedBy": [], "createdAt": "2026-10-01T06:58:00Z", "updatedAt": "2026-10-02T16:25:17Z", "preview": "Placement proposal §2 (owner rulings 2026-09-30 §6; process: review → resolve issues → staged PRs). Parent of PR0/PR1 is `9umr`; PR2–PR9 sit under the separation epic because D4 (owner, 2026-10-01, option 3) interleaves them with the split stages into one ordered sequence. Every PR: references move with the subject; generated trees regenerated, never hand-edited; `beans/**` not rewritten; kg-qa si", "file": "beans/defs/folio-assistant-63wl--placement-pr3-bpmn-and-dmn-move-up-to-their-owners.md" }, { "id": "folio-assistant-65oe", "title": "library:viz is not reproducible across environments: refScan.filesRead embeds a count of the files on disk", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T15:40:27Z", "updatedAt": "2026-10-04T06:08:20Z", "preview": "## The defect\n\n`cat-harness/docs/assets/library/index.json` ends with\n\n \"refScan\": { \"filesRead\": 2663, \"unreadable\": [] }\n\n`filesRead` is a COUNT OF THE FILES THE SCAN READ ON DISK. It is therefore a\nproperty of the checkout, not of the library, and it is committed. Any two\nenvironments whose trees differ by even one scanned file disagree, so\n`library:viz:check` reports stale whenever the last", "file": "beans/defs/folio-assistant-65oe--libraryviz-is-not-reproducible-across-environments.md" }, { "id": "folio-assistant-65zi", "title": "MIGRATE req:* statements to carry successCriteria, then make the field required (#2405 decision 3)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-10-07T11:04:33Z", "updatedAt": "2026-10-09T13:07:00Z", "preview": "Issue: https://github.com/litlfred/folio-assistant/issues/2405 — owner decision 3 (https://github.com/litlfred/folio-assistant/issues/2405#issuecomment-6035163987): successCriteria was added OPTIONAL on RequirementStatementFields (bootstrap-tools/schemas/requirement.ts, PR litlfred/folio-assistant#2407); check:requirements WARNS on a statement without it. Measured on the branch: 31 statements acro", "file": "beans/defs/folio-assistant-65zi--migrate-req-statements-to-carry-successcriteria-th.md" }, { "id": "folio-assistant-676g", "title": "kg:validate cannot own a NESTED instance's path, so 15 instances' committed QA sidecars are consumer-validated by nothing", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-27T16:50:03Z", "updatedAt": "2026-10-02T06:44:57Z", "preview": "## What was measured\n\n`kg:validate` resolves declared directories from ONE instance root, so a sidecar\nin a nested instance is refused rather than validated. Measured 2026-09-27, from\nthe repository root:\n\n bun run cat-harness/scripts/kg-validate.ts cat-harness/test/results/kg-qa/processes/adjudication.kg-qa.json\n ✓ … [qa]\n\n bun run cat-harness/scripts/kg-validate.ts bootstrap/test/resu", "file": "beans/defs/folio-assistant-676g--kgvalidate-cannot-own-a-nested-instances-path-so-1.md" }, { "id": "folio-assistant-680p", "title": "IG render: load from the KG client-side to cut .html bloat", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T16:11:22Z", "updatedAt": "2026-10-07T20:30:00Z", "preview": "Owner, 2026-10-01, verbatim: \"note, other goals of jsutthedoc rendering is to reduce the .html bloat.... lots of it can be loaded client side from the KG\".\n\nThe Publisher writes every view as a full static HTML page, though its own DAK view pages already fetch their file in the browser. A just-the-docs page should carry identity and layout, and fetch repeated or bulky content at view time from the", "file": "beans/defs/folio-assistant-680p--ig-render-load-from-the-kg-client-side-to-cut-html.md" }, { "id": "folio-assistant-68k7", "title": "resolveDirectories climbs the checkout's PARENT: is that intended for the ROOT instance?", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T19:48:10Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "**Owner ruling, 2026-09-30: leave open until the call site is named.**\n\n## Why this is its own bean\n\nIt was Box 3 of `8zsb`, and **`8zsb` is `completed`**. A completed bean's\nunchecked box is not reachable by `beans list`, so the question was on its way\nto being invisible — which is `8zsb`'s own lesson (*a finding needs somewhere\nto live that nobody has to already know about*) turned on the bean t", "file": "beans/defs/folio-assistant-68k7--resolvedirectories-climbs-the-checkouts-parent-is.md" }, { "id": "folio-assistant-68op", "title": "GATE NAME DESCRIBES ONE OF TWO FAILURE MODES: 'viewer pages keep the navbar they had' goes red when the sidecar is ABSENT, and the same run says 0 pages regressed", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T01:08:12Z", "updatedAt": "2026-10-07T11:50:39Z", "preview": "The step name in `code-quality-gates.yml` names the REGRESSION branch of\n`check:viewer-nav`. The check has two failure modes, and the other one -\nthe sidecar is absent, so the audit cannot see its subject - fails under\nthe same name. A reader then goes looking for a page that lost its rail,\nand the same run tells them none did.\n\n## Measured 2026-10-03 (on PR #1819, heads c975c7da7b9 and 28746eb6a8", "file": "beans/defs/folio-assistant-68op--gate-name-describes-one-of-two-failure-modes-viewe.md" }, { "id": "folio-assistant-6b8u", "title": "EXTRACTION: MD_MIN_TEXT_LEN is a threshold in CHARACTERS, so the same table cell is translatable in Arabic and not in English", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T08:56:25Z", "updatedAt": "2026-09-27T08:06:12Z", "preview": "Found while deriving `.po` catalogues for #206 (bean `tbdg`). Measured\n2026-09-26 on `main` + this branch.\n\n`MD_MIN_TEXT_LEN = 3` in `content/pipeline/pot-extract.ts` gates every one of the\nfour emission paths (heading, paragraph, list-item, blockquote, table-cell). It is\na minimum in **characters**, used as a proxy for \"is there translatable text\nhere\". A character count is not script-neutral, so", "file": "beans/defs/folio-assistant-6b8u--extraction-md-min-text-len-is-a-threshold-in-chara.md" }, { "id": "folio-assistant-6bhf", "title": "SECURITY SUB-KG: an external identifier reaches join() unvalidated in three HTTP handlers, and the containment helper that fixes it already exists in one script", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-25T16:38:20Z", "updatedAt": "2026-09-26T14:31:39Z", "preview": "Owner, 2026-09-25, verbatim:\n\n> add lots of precautions/double checking/fails safe on rm -fr... is variable\n> there? injection hazard. that is needed on ALL I/O. bean as ceybeesquity sub\n> KG in tools. first is injection et. related skills already man y in folio and\n> smart-base. consolidate.\n\nOwner chose scope **2 + 3**: consolidate the scattered pieces **and** audit every\nI/O site. This bean is ", "file": "beans/defs/folio-assistant-6bhf--security-sub-kg-an-external-identifier-reaches-joi.md" }, { "id": "folio-assistant-6brk", "title": "FLAKE: check-declaration-filename's corpus scan runs 4.2s against a 5000ms test budget on a clean main — any corpus growth tips it", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T06:21:41Z", "updatedAt": "2026-09-24T19:53:28Z", "preview": "`check-declaration-filename.test.ts` scans the whole markdown and workflow\ncorpus. **Measured 2026-09-24, directly, three runs per tree:**\n\n| tree | ms per `checkDeclarationFilename()` |\n|---|---|\n| clean `main` | **4476, 4198, 3927** |\n| `main` + ~150 lines of markdown (PR #1188) | **4927, 4216, 4249** |\n\nThe test budget is **5000 ms**. So on a clean `main`, before anybody adds\nanything, the scan", "file": "beans/defs/folio-assistant-6brk--flake-check-declaration-filenames-corpus-scan-runs.md" }, { "id": "folio-assistant-6dvy", "title": "Witnessed values (:val): prose bug, just-the-docs substitution, namespacing by source", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T10:08:52Z", "updatedAt": "2026-09-30T15:42:28Z", "preview": "Owner 2026-09-30: fix the platform; first analyze consistency with the just-the-docs pipeline, and how to namespace values by source (paper, dataset e.g. CODATA with provenance via ingestion, FHIR IG metadata).\n\n## Done when\n- [x] prose :val resolves names with underscores (render-latex), tested and calibrated\n- [x] consistency analysis across PDF / just-the-docs / blueprint / FHIR IG, recorded (b", "file": "beans/defs/folio-assistant-6dvy--witnessed-values-val-prose-bug-just-the-docs-subst.md" }, { "id": "folio-assistant-6eiw", "title": "Library titles for non-PDF entries (pptx deck, CODATA table)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T12:55:31Z", "updatedAt": "2026-10-12T09:12:00Z", "preview": "## Why\n\nBean `w6fu` (PR #1849) gave PDF library entries a verified title: a title is\nused only when two independent sources agree, or when an editorial\n`title_correction` exists. Two entries are not PDFs, so that resolver never\nreaches them, and they still show placeholder titles:\n\n| entry | format | what shows now | evidence in the file |\n|---|---|---|---|\n| `cat-harness/library/kg-folio-asst-202", "file": "beans/defs/folio-assistant-6eiw--library-titles-for-non-pdf-entries-pptx-deck-codat.md" }, { "id": "folio-assistant-6g6o", "title": "CI: PR-time Liquid undefined-variable and post-merge regen checks block before merge", "status": "in-progress", "type": "task", "priority": "normal", "parent": "folio-assistant-ksek", "blocking": [], "blockedBy": [], "createdAt": "2026-10-23T08:43:02Z", "updatedAt": "2026-10-23T08:43:02Z", "preview": "Rule: a check that today runs only after merge gets a PR-time half wherever it can run on a PR, counted by merge-guard. PR: https://github.com/litlfred/folio-assistant/pull/2633 (branch claude/pr-time-postmerge-checks, head 7d2ea91). What it does: liquid-undefined.yml is called (workflow_call) as a job of Feature Staging's own run after stage hands off the Jekyll source, so a red check makes the P", "file": "beans/defs/folio-assistant-6g6o--ci-pr-time-liquid-undefined-variable-and-post-merg.md" }, { "id": "folio-assistant-6gme", "title": "Remaining single-instance kinds have no viewer: scenarios, policies, UML, test plan, tool releases, vocab mapping, voice vendors, waivers, substrate snapshots, basic CDN site, Lean packages, bootstrap-tools skills", "status": "in-progress", "type": "feature", "priority": "normal", "parent": "folio-assistant-46ob", "blocking": [], "blockedBy": [], "createdAt": "2026-10-15T16:53:28Z", "updatedAt": "2026-10-23T08:00:00Z", "preview": "## Parent\n\nfolio-assistant-46ob\n\nMeasured live: Scenarios ×3, and one row each for Policies, UML, Test plan, Tool releases, Vocab mapping, Voice vendors, Waivers, Substrate snapshots, Basic CDN site (C@T Harness), Lean packages (folio-assistant-sci; a doc-gen4 viewer exists as the `lean-html-docs` skill), Skills (Bootstrap tools only — other harnesses' skills link).\n\nTriage each: link an existing ", "file": "beans/defs/folio-assistant-6gme--remaining-single-instance-kinds-have-no-viewer-sce.md" }, { "id": "folio-assistant-6h47", "title": "RENDER STAGE 3: the dynamic-state export has a position but no filename — the owner left it open on purpose", "status": "todo", "type": "task", "priority": "normal", "parent": "folio-assistant-yj32", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T19:59:52Z", "updatedAt": "2026-10-03T07:43:33Z", "preview": "Owner, 2026-09-20, settling where the README render sits in the pipeline:\n\n> dynamic content then hets get json/jsonld export avaiable under\n> cat-harness/state.jsonld or so (whatever matches the dyanmic state\n> vsualizesrs)\n\n**\"or so (whatever matches …)\" is an explicit uncertainty, and it was not\nclosed by guessing.** `scripts/render-pipeline.ts` declares the step\n(`kg-dynamic`), fixes its POSIT", "file": "beans/defs/folio-assistant-6h47--render-stage-3-the-dynamic-state-export-has-a-posi.md" }, { "id": "folio-assistant-6hf2", "title": "A: data-modelling guidance — dependents point at the general node; references are typed KG refs", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T19:53:34Z", "updatedAt": "2026-09-23T20:12:11Z", "preview": "## Done when\n- [x] data-modelling.md states both rules with the owner's examples and the existing typed-ref helpers\n- [x] skill prose that names its callers is removed (confirm-harness, root-readme, discussion, log-message, bootstrap package manifest)\n- [x] gates green, PR open\n\n## Summary of Changes\n\nMerged in #1172 (aabdc87): data-modelling.md steps 8 (dependents point at the general node; the g", "file": "beans/defs/folio-assistant-6hf2--a-data-modelling-guidance-dependents-point-at-the.md" }, { "id": "folio-assistant-6lb8", "title": "FOLIO BOARD: a resizable Miro-style board, notes that move and attach, and semantic zoom to avatars", "status": "in-progress", "type": "epic", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T10:37:35Z", "updatedAt": "2026-10-05T04:57:13Z", "preview": "## The ask, owner 2026-09-20 (verbatim, in the order it arrived)\n\n> landing folio should really be resizable, switchiing over to content avatars\n> if content no longer lgeible. liek miro board. content vieweing panes are\n> resizzble. but ALWAYS collapsable to linearly rendablee/just the docs. the\n> todos/miro board is a static content overlays, so new content type that sits\n> under todos/ but scem", "file": "beans/defs/folio-assistant-6lb8--folio-board-a-resizable-miro-style-board-notes-tha.md" }, { "id": "folio-assistant-6lre", "title": "CI WATCH READS A PARTIAL CHECK SET AS GREEN: 1 of 13 registered, verdict PASS — ask the check SUITES, not only the runs", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T16:15:00Z", "updatedAt": "2026-10-06T14:57:00Z", "preview": "Found while using the tool, not while reading it. Sibling of `2c2b`, which\nshipped `ci:watch` for issue #1624.\n\n## The defect, measured\n\n`bun run cat ci:watch 29b10a68923` at 16:13:58:\n\n```\n16:13:58 29b10a68923 PASS — 1 check(s) completed clean\n```\n\nexit **0**. At that instant `get_check_runs` on the same commit returned\n`total_count: 1` — the `.jsonld siblings` job — and **nothing else had been", "file": "beans/defs/folio-assistant-6lre--ci-watch-reads-a-partial-check-set-as-green-1-of-1.md" }, { "id": "folio-assistant-6mk7", "title": "A refused packages.fhir.org did not lead the agent to fhir-cache-seed-npm: symptom-to-Tool lookup is missing, and the mirror the skill names does not exist", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T06:31:37Z", "updatedAt": "2026-10-06T16:00:00Z", "preview": "Owner, 2026-10-06: *\"why didnt you know about the tool already? what KG process broke down?\"*\n\n## What happened\nSession https://claude.ai/code/session_01EcBv3uwKYcnNbCC6BcPG92 needed SUSHI clean on three IG repositories (owner preference). SUSHI failed: packages.fhir.org is refused by the egress proxy. The agent told the owner \"SUSHI can't run in this container\" and moved on. The owner pointed at ", "file": "beans/defs/folio-assistant-6mk7--a-refused-packagesfhirorg-did-not-lead-the-agent-t.md" }, { "id": "folio-assistant-6o1z", "title": "crdm-detect corpus: #187 and #199 are the same artefact, labelled opposite", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-ahvw", "blocking": [ "folio-assistant-vjbl" ], "blockedBy": [], "createdAt": "2026-09-21T16:40:00Z", "updatedAt": "2026-09-21T16:40:00Z", "preview": "**A specific, checkable contradiction in the eval corpus — not a judgement call\nabout intent.** `vjbl` already names #187 as a hard case *\"turning on intent\nrather than wording\"*. Measured while working `9gtc`, it is sharper than that:\n\n| issue | label | what it is |\n|---|---|---|\n| [#187](https://github.com/litlfred/folio-assistant/issues/187) | `isFeature: false` — *\"a documentation deliverable ", "file": "beans/defs/folio-assistant-6o1z--crdm-detect-corpus-187-and-199-are-the-same-artefa.md" }, { "id": "folio-assistant-6pfo", "title": "Staging previews on gh-pages carry metadata that exists only at runtime — publish it as a KG graph", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T12:28:52Z", "updatedAt": "2026-10-07T17:15:00Z", "preview": "Owner, 2026-09-19: *\"metadata on ghpages staging should be avalabale in KG\"*.\n\nQueued, not started — raised mid-turn while the preview cleanup was running,\nand the standing instruction is to queue rather than pivot.\n\n## The ask, as I read it\n\nEvery staging preview under `STAGING/` on `gh-pages` has facts attached\nto it: which branch it was built from, which commit, which pull request and\niss", "file": "beans/defs/folio-assistant-6pfo--staging-previews-on-gh-pages-carry-metadata-that-e.md" }, { "id": "folio-assistant-6ptx", "title": "COORDINATION: eight sessions woke into the same 2435-commit gap and all eight re-surveyed it independently", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-25T15:44:18Z", "updatedAt": "2026-09-26T17:32:56Z", "preview": "## Measured 2026-09-25, from the session API\n\n`list_sessions` returned 30 sessions, of which **eight were RUNNING at the same\nminute** (15:38–15:39Z). Their own `task_summary` fields, verbatim:\n\n```\nrebasing from main; 2369 commits (PR #477→#1306); facts before plan\nreviewing PR #542 impact; fast-forward reset blocked; surveying origin/main\nre-fetching main after date change; will propose next ste", "file": "beans/defs/folio-assistant-6ptx--coordination-eight-sessions-woke-into-the-same-243.md" }, { "id": "folio-assistant-6qk5", "title": "QA REVIEW MODE: translation QA joins the audited review record under test/results", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T08:55:36Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "From [#363](https://github.com/litlfred/folio-assistant/issues/363): \"qa review\nmode: test results go under things like test/results and same for trnslations\nb/c Q&A is part of an audited review process\".\n\n## Half of this has landed; the bean is the other half\n\nMeasured 2026-09-19 against `harness.json`: `test/results/` **is** declared, as\nthe `qa` graph, holding three kinds distinguished by their", "file": "beans/defs/folio-assistant-6qk5--qa-review-mode-translation-qa-joins-the-audited-re.md" }, { "id": "folio-assistant-6rmv", "title": "B10e (#1168): instance references are owner/repo everywhere", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T14:59:54Z", "updatedAt": "2026-09-30T19:28:44Z", "preview": "Owner 2026-09-30: owner/repo everywhere (over the recommended instance names). 137 instance values (136 names, 1 owner/repo) across AssetSource.instance, ThemeRef, voices, role refs, session-context, tool-invocation, site-indexes. Needs the instance→repo mapping from each .json.\n## Done when\n- [ ] every instance-reference field is RepoFullName-typed and every value converted\n\n\n\n## Scope ", "file": "beans/defs/folio-assistant-6rmv--b10e-1168-instance-references-are-ownerrepo-everyw.md" }, { "id": "folio-assistant-6xaz", "title": "pdf-structure infers a TOC from a worked EXAMPLE and ships it as the document's own structure", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-0lmb", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T00:12:09Z", "updatedAt": "2026-10-10T16:40:52Z", "preview": "Measured 2026-09-19 ingesting `uploads/WPR-RDO-2020-003-eng.pdf` (WHO WPRO 'Publication and Information Products Style Guide', 33pp, no embedded outline).\n\n`scripts/pdf-structure.py` inferred 13 sections and wrote 11 of them named after **a different publication**: 'FACILITY LEVEL: Improving hospital planning and management', '1.1 Accountability', '2.1 Goal – Hospitals as a path to UHC'. Those com", "file": "beans/defs/folio-assistant-6xaz--pdf-structure-infers-a-toc-from-a-worked-example-a.md" }, { "id": "folio-assistant-6xdf", "title": "PUBLIC COMMENT: a PR body that merely MENTIONS a change-set id re-links an already-incorporated change-set to it, and the record writer drops $schema", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-q4jm", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T14:36:04Z", "updatedAt": "2026-10-07T09:44:54Z", "preview": "Measured 2026-10-06 on litlfred/smart-ra#24 (a submodule-pin staging PR whose body named CS-236 and CS-237 in prose). The public-comment workflow (pull_request_target, platform pinned at 7a2ff7e) committed 5f4ce41 to smart-ra main: both change-sets' pr moved from #9 (the merged PR that made the change) to #24, history gained 'PR #24 is making the change', and both files LOST their \"$schema\": \"chan", "file": "beans/defs/folio-assistant-6xdf--public-comment-a-pr-body-that-merely-mentions-a-ch.md" }, { "id": "folio-assistant-6ztw", "title": "Node kinds: generic pages at ////[/[]] (#2195 PR 2)", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-10-05T22:30:10Z", "updatedAt": "2026-10-06T05:52:19Z", "preview": "Issue #2195 PR 2. Generic dashboard per kind (subclasses included, filters by kind and harness, status tiles from enum fields), per-harness dashboard, per-node page; en only; generated by node-kind:pages, checked in CI. Next: PR 3 change-set override + pages-resolve QA; folio builds (smart-ra) run the generator.\n\n\n## Summary of Changes\n\nMerged as #2242 (`3c54172` on main); PR 3 (#2243, `7a2ff7e`) ", "file": "beans/defs/folio-assistant-6ztw--node-kinds-generic-pages-at-localedeclaringkindhar.md" }, { "id": "folio-assistant-70c7", "title": "docs/assets/qa was undeclared: 134 committed files no consumer scans", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T06:33:39Z", "updatedAt": "2026-09-23T19:25:46Z", "preview": "## Summary of Changes\n\nClosed 2026-09-23 **on evidence, not authorship**, in the owner's \"go through remaining beans\" sweep. A read-only check against `main` called it landed, and it was re-verified before closing:\n\nThe witnesses now live in the declared `qa` graph at `cat-harness/test/results/witnesses/`. `cat-harness/docs/assets/qa/` holds only `index.json`, which `gen-docs-pages.ts` writes, and", "file": "beans/defs/folio-assistant-70c7--docs-assets-qa-was-undeclared-134-committed-files-.md" }, { "id": "folio-assistant-70lx", "title": "Separation stage 1a: stage cat-harness-tools/ as a sibling instance and git mv the unambiguous code", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-iirv", "blocking": [ "folio-assistant-iai8", "folio-assistant-w2gr" ], "blockedBy": [ "folio-assistant-pyds" ], "createdAt": "2026-10-01T06:58:00Z", "updatedAt": "2026-10-16T06:55:00Z", "preview": "Stage 1a of the split plan: stage `cat-harness-tools/` as a sibling instance in this repo and `git mv` the unambiguous code (≈1,340 renames). Owner D1 (2026-10-01): ALL cat-harness code moves; dependents' code may import cat-harness-tools. D4 (2026-10-01): 1a runs NOW, before placement PR2.\n\nPlans (session scratchpad, 2026-10-01; to be committed with stage 0): `cat-harness-split-plan.md` (stages 0", "file": "beans/defs/folio-assistant-70lx--separation-stage-1a-stage-cat-harness-tools-as-a-s.md" }, { "id": "folio-assistant-70zt", "title": "SVG rendering of DIIG Fig. 1.1.1 (seven phases of planning and implementing a digital health enterprise)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-qvxh", "blocking": [], "blockedBy": [ "folio-assistant-5ki9" ], "createdAt": "2026-10-03T10:55:02Z", "updatedAt": "2026-10-04T12:41:46Z", "preview": "Owner, 2026-10-03 (issue #1984, relayed from session_015Q15h1fg2Hh9MJXfAqr4h7): \"use one DIIG seven phase figure as source. need SVG rendering maybe as bean\".\n\nThe figure is DIIG §1.1 Fig. 1.1.1, smart-base/library/9789240010567-eng, PDF p.17 (the phase list is on p.16). vector-labels.json records 82 drawings on that spread, and no raster figure. It is vector-drawn, so its home is the vector-figur", "file": "beans/defs/folio-assistant-70zt--svg-rendering-of-diig-fig-111-seven-phases-of-plan.md" }, { "id": "folio-assistant-72a8", "title": "5hox readers: the six gates + tests that still read committed test/results/ on main after #1801", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T09:07:28Z", "updatedAt": "2026-10-07T17:47:00Z", "preview": "Measured 2026-10-04 on main 12b916e with the qa:verify-moved inventory removed (1,223 files): 10 of 228 gates red (issue #1763 comment 5978225547). Owner, 2026-10-04: \"lets move the qa-reports off main\", then \"Yes, do the readers\". Pattern: #1801's compute-and-judge (a gate's own results absent -> recompute and judge), or fetch from cat/cat-harness/qa-reports. Agent/human verdicts must never be si", "file": "beans/defs/folio-assistant-72a8--5hox-readers-the-six-gates-tests-that-still-read-c.md" }, { "id": "folio-assistant-72gk", "title": "todos visualiser: 5 wireframe findings", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T10:36:15Z", "updatedAt": "2026-10-09T12:52:00Z", "preview": "Findings from the as-is wireframe `cat-harness/docs/wireframes/todos/` (intent.md, as-is.html, checks/), observed at 1280×800 and 390×844. Verbatim from its `## Findings`; a finding tagged → is also covered by that cross-cutting bug.\n\n1. **The items are not on the page.** `assets/todos/index.json` carries 3 items, each with a summary, status, priority, target page and `viewHref`/`editHref`: \"The h", "file": "beans/defs/folio-assistant-72gk--todos-visualiser-5-wireframe-findings.md" }, { "id": "folio-assistant-750a", "title": "Work-state kinds have no viewer: bean defs, bean notes, workflow state, memory, sweep marks, boards, merge queue, surveys, todo feedback/items, interaction", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-46ob", "blocking": [], "blockedBy": [], "createdAt": "2026-10-15T16:53:28Z", "updatedAt": "2026-10-21T17:27:00Z", "preview": "## Parent\n\nfolio-assistant-46ob\n\nMeasured live: Bean definitions ×3, Bean notes ×3, Workflow state ×3, Memory ×2, Sweep marks ×1, Boards ×1, Board positions ×1, Merge queue ×1, Session surveys ×1, Todo feedback ×1, Todo items ×1, Interaction ×1.\n\nSeveral are parts of a graph that already has a viewer (bean defs and notes are inside Beans; workflow state is the BPMN instances the Processes viewer c", "file": "beans/defs/folio-assistant-750a--work-state-kinds-have-no-viewer-bean-defs-bean-not.md" }, { "id": "folio-assistant-792f", "title": "state:mount rewrites the .gitignore state-mounts block from STALE layer mounts, dropping graphs the pinned layers declare (seen: /merge-reviews/**)", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-22T15:00:02Z", "updatedAt": "2026-10-22T17:07:00Z", "preview": "`bun run cat state:mount -- --id beans` rewrites the `# BEGIN/END state mounts` block of the root `.gitignore` and drops `/merge-reviews/**`. This was seen twice on 2026-10-22, both times in a checkout at main `6be6532`'s ancestors. The diff each time was:\n\n```\n-/merge-reviews/**\n```\n\nEvery agent that mounts beans then has a dirty tree it did not mean to make, and the stop hook asks it to commit. ", "file": "beans/defs/folio-assistant-792f--statemount-id-one-drops-merge-reviews-from-the-git.md" }, { "id": "folio-assistant-79t3", "title": "A repo's type set is the markers it carries, closed under the dependency tree", "status": "todo", "type": "task", "priority": "normal", "parent": "folio-assistant-vke6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-18T21:28:21Z", "updatedAt": "2026-09-23T06:15:45Z", "preview": "## The idea, as stated 2026-09-18\n\n**A repo declares itself an instance of a content type by carrying a file named\nafter that type.** The filename is the assertion; the file's JSON-LD `@type` is\nthe resolvable reference to what that type *means*. Several markers coexist, so\na repo is a **set** of types, not one: `smart-base` is a DAK *and* a SUSHI\nproject. Each marker stays owned by whoever define", "file": "beans/defs/folio-assistant-79t3--a-repos-type-set-is-the-markers-it-carries-closed.md" }, { "id": "folio-assistant-7bg9", "title": "INGESTION: a licence check subprocess, and the marker coverage the ingest family is missing", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T15:25:27Z", "updatedAt": "2026-09-30T10:26:57Z", "preview": "Owner, 2026-09-20: *\"liscence check task (subprocess) is part of document/asset\ningestion pipeline\"*.\n\nQueued rather than started, per the standing instruction to queue new topics in\nparallel and not pivot.\n\n## The gap is real, and the shape already fits\n\nMeasured 2026-09-20 against `origin/main` at `dc78e7ccf`:\n\n cat-harness/processes/document-ingestion.bpmn\n \"Document ingestion — uploads", "file": "beans/defs/folio-assistant-7bg9--ingestion-a-licence-check-subprocess-and-the-marke.md" }, { "id": "folio-assistant-7deg", "title": "LIBRARIAN: an avatar for knowledge content, and Dublin Core introduced in folio-assist-core", "status": "completed", "type": "feature", "priority": "", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T06:23:14Z", "updatedAt": "2026-10-10T08:00:00Z", "preview": "## The ask, owner 2026-09-20 (verbatim)\n\n> also need avator of librarian for knwoledge content specific stuff in\n> folio-asst-core where dubln core skill should be introduced for first time.\n\n## Two things, and they are separable\n\n1. **A librarian avatar** for knowledge-content-specific material in\n `folio-assist-core`.\n2. **A Dublin Core skill**, introduced there for the first time.\n\n## What ex", "file": "beans/defs/folio-assistant-7deg--librarian-an-avatar-for-knowledge-content-and-dubl.md" }, { "id": "folio-assistant-7dek", "title": "render-kg-to-cdn: one general publish-to-CDN step; gh-pages is one Tool with its own subprocess", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-iirv", "blocking": [], "blockedBy": [ "folio-assistant-hx65" ], "createdAt": "2026-10-01T06:58:00Z", "updatedAt": "2026-10-06T19:30:00Z", "preview": "Owner, 2026-09-30: gh-pages is \"one specific tool of general 'publish to CDN'\"; render a KG (or a list of subgraphs) to a CDN at a publication root URL, independent of staging vs release; output = push status + message; \"tools can describe their own specific subprocesses if needed to not bog down general skills\" (placement ruling 6). Owner, 2026-10-01: a gh-pages branch must exist before Pages can", "file": "beans/defs/folio-assistant-7dek--render-kg-to-cdn-one-general-publish-to-cdn-step-g.md" }, { "id": "folio-assistant-7e59", "title": "INGEST: decision-making methodologies — skill takes decision context as input, outputs ranked applicable methods with criteria and rationale. Source: qou bd0c2cb7 (3 arxiv PDFs: 2508.21620 probabilistic/bandits, 2509.06388 MCDM/AHP/SAW, 2607.20636 sequential/social). Covers all methodology families with when-to-use criteria.", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-slw1", "blocking": [], "blockedBy": [], "createdAt": "2026-09-25T15:38:03Z", "updatedAt": "2026-10-07T17:37:00Z", "preview": "## Body added by a gate fix, not by this bean's author\n\n**This is not the author's scoping.** `check:bean-bodies` was red on `main`\nwith `[empty-body]: \"todo\" with front matter and nothing else — a sibling\nreading the store learns nothing about it`, and `check:bean-parents` was red\nbecause this bean carried no `parent:`. Both were fixed here as part of\nclearing `main`'s red gates, on the owner's \"", "file": "beans/defs/folio-assistant-7e59--ingest-decision-making-methodologies-skill-takes-d.md" }, { "id": "folio-assistant-7eak", "title": "Dublin Core renderings: DC XML + JSON(-LD) per record, as a skill and tool in the rendering pipeline, published to gh-pages; who-iris links to both", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-7deg", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T08:54:51Z", "updatedAt": "2026-10-09T17:42:54Z", "preview": "## The ask, owner 2026-09-30 (verbatim)\n\n> do we render the proper xml for dublin core? it should be in rendering ppiple as skill and tool. i tihnk. and then pushed to gh-pages\n> that is in cat-harness... bean up for now\n> (and who-iris should link to json and xml renderings)\n\n## What is true today (measured 2026-09-30, not yet investigated in depth)\n\n- Dublin Core terms are used as JSON-LD pr", "file": "beans/defs/folio-assistant-7eak--dublin-core-renderings-dc-xml-json-ld-per-record-a.md" }, { "id": "folio-assistant-7fu5", "title": "CI: fold the four one-second jobs into one — fewer runner requests per run", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-hfag", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T16:14:26Z", "updatedAt": "2026-10-04T05:33:46Z", "preview": "Measured on main run 37133247484 (2026-10-03, #2010 merge): 21.7 min wall for a run whose longest job executes 174 s. Runner WAIT dominates: first job queued 4.5 min; the two 1-s aggregator jobs waited 9.6 and 3.8 min. Each run asks for 18 runners. Four jobs do ~1 s of work after a ~13 s checkout: lean-bare-import, python-imports, rust-wildcard, dependency-advisories.\n\nOwner chose option 1 (2026-1", "file": "beans/defs/folio-assistant-7fu5--ci-fold-the-four-one-second-jobs-into-one-fewer-ru.md" }, { "id": "folio-assistant-7h1c", "title": "publish-verify scope misses our own SKOS code-lists document", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T05:40:40Z", "updatedAt": "2026-09-24T11:55:19Z", "preview": "Found 2026-09-24 while working `wg7r`. `publish-verify`'s `isOurs` counts a JSON-LD document as ours only when its `@context` names our content context or binds one of our namespaces. `-code-lists.jsonld` (glossary-export) binds only `skos`, `dcterms`, `owl` and `rdf`, so the verifier counts it **out of scope (not ours)**, even though its `@id` sits under our published base and we author eve", "file": "beans/defs/folio-assistant-7h1c--publish-verify-scope-misses-our-own-skos-code-list.md" }, { "id": "folio-assistant-7h3u", "title": "smart-trust replica wears folio-assistant's chrome, not the WHO IG's branding", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T20:04:11Z", "updatedAt": "2026-10-07T17:37:00Z", "preview": "**Owner, 2026-09-30:** *\"does not make css/branding/style of\nworldhealthorganization.github.io/smart-trust from\ngithub.com/worldhealthorganization/smart-trust\"*\n\nObserved at : dark\nbackground, folio-assistant's left rail and its purple **▾ Folio** handle. The\npublished WHO IG it replicates uses the HL7 IG-publisher template with WHO\nbranding", "file": "beans/defs/folio-assistant-7h3u--smart-trust-replica-wears-folio-assistants-chrome.md" }, { "id": "folio-assistant-7how", "title": "REGEN READS A STALE QA TREE: generators that read gitignored */test/results write wrong pages unless qa:working-copy ran first", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T09:15:31Z", "updatedAt": "2026-10-07T17:47:00Z", "preview": "Measured 2026-10-06 on #2267 (coordinator) and reported independently by session A, from #2268.\n\n**What happens.** Since #2080 (5hox), the QA results tree is computed and not committed. Some generators read it from disk, at least `uml:overview` and `readme:subgraphs`. A container's local copy can be partial or stale. On #2267 only `kg-qa` was present, so `bun run cat regen` rewrote the QA overview", "file": "beans/defs/folio-assistant-7how--regen-reads-a-stale-qa-tree-generators-that-read-g.md" }, { "id": "folio-assistant-7jdm", "title": "Merge refused: #2055 owed CI not green on its head", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-nok9", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T14:30:41Z", "updatedAt": "2026-10-07T14:55:47Z", "preview": "PR #2055, handed back by the Merge Manager on 2026-10-04 (BACKFILLED at 14:50Z: the hand-back was made as a PR comment and a queue entry only, without this bean. Owner: \"Backfill + enforce\").\n\n## Refused\n- `merge:steward` at 14:30Z: Rule_HeadNotGreen: conflict declared, own CI missing-required.\n- Queue entry (`beans/queue/litlfred--folio-assistant--2055.json`, status `waiting-on-author`): Covered ", "file": "beans/defs/folio-assistant-7jdm--merge-refused-2055-owed-ci-not-green-on-its-head.md" }, { "id": "folio-assistant-7ji4", "title": "id-lookup page has no rail: find why and fix", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-9rq1", "blocking": [], "blockedBy": [], "createdAt": "2026-10-05T15:23:55Z", "updatedAt": "2026-10-07T11:50:40Z", "preview": "Left over from lnoy (#2185): of the 95 navbar-less pages, id-lookup is the one not covered by the thin-page linked rail. Not investigated.\n\n_2026-10-06T19:13:41Z_ — Claimed by claude/7ji4-id-lookup-rail — pushed to main so sibling sessions see it before this branch has a PR (bean 35nj).\n\n## Evidence\n- Root cause:\n 1. `cat-harness-tools/id-lookup/index.html` was missing ` workflows/ is declared under the id cat-harness-workflows, content should be\n> at graph cat-harness/workflows except the bare minimum declarations (what is\n> bpmn, how use) under bootstrap/,\n> cat-harness/workflows is for more elboration on use of bpmn and such\n> cat-harness/workflows/state should describe beans and todos", "file": "beans/defs/folio-assistant-7po1--workflows-bootstrap-keeps-the-bare-minimum-cat-har.md" }, { "id": "folio-assistant-7pp6", "title": "A column-0 pseudo-tag in a published markdown source is caught only AFTER merge — five hand-fixes and no gate", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T06:29:56Z", "updatedAt": "2026-10-07T11:50:39Z", "preview": "## The class, and why a sixth hand-fix is the wrong answer\n\nA markdown source line whose first character is `<` followed by a non-HTML name\nmakes kramdown open a **raw HTML block**. With no closing tag the rest of the\npage stays raw, so headings, lists and tables after it render as literal text.\nThe usual cause is an inline code span **wrapped across a line break**, leaving\nits continuation at col", "file": "beans/defs/folio-assistant-7pp6--a-column-0-pseudo-tag-in-a-published-markdown-sour.md" }, { "id": "folio-assistant-7ppv", "title": "resolveCanonicalLean is O(refs x tree) without a cache, and its two index builders walk .lake/ while listPackageLeanFiles excludes it", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-27T09:37:16Z", "updatedAt": "2026-10-07T17:18:00Z", "preview": "Found 2026-09-27 while moving the resolver from `content/pipeline/qa-utils.ts`\ninto `content/pipeline/lean-formal-ref.ts` (PR #1465). **The code is unchanged**\n— it moved verbatim — so this is a pre-existing cost, not a regression, and it\nis filed because it was measured rather than because the move introduced it.\n\n## What is measured\n\n`resolveCanonicalLean(ref, repoRoot, cache?)` takes an OPTIONA", "file": "beans/defs/folio-assistant-7ppv--resolvecanonicallean-is-orefs-x-tree-without-a-cac.md" }, { "id": "folio-assistant-7pvn", "title": "folio-assistant-core imports cat-harness (and cat-harness-tools) as packages, not by ../../", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-f6gq", "blocking": [], "blockedBy": [], "createdAt": "2026-10-09T15:29:13Z", "updatedAt": "2026-10-14T16:30:00Z", "preview": "Second layer of f6gq; needs pxtk. cat-harness needs `exports` (or deep-import-safe layout) and a pin agreeing with core's index.config.json.\n\n## Done when\n- [x] core's 228 cross-repository imports become package imports — 260 measured and rewritten in litlfred/folio-assistant-core#80; merged as core a2342ad, on cat-harness-tools#198 (f246be6, exports derived from declared code subgraphs). Downstre", "file": "beans/defs/folio-assistant-7pvn--folio-assistant-core-imports-cat-harness-as-a-package.md" }, { "id": "folio-assistant-7s52", "title": "staging-review hands a reader URLs the AGENT cannot open — but the publish ref is a git branch", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T06:33:57Z", "updatedAt": "2026-09-30T14:44:36Z", "preview": "Issue: https://github.com/litlfred/folio-assistant/issues/849\n\n`staging-review` already owns the publish ref, and already names the failure:\nbean `rptk`, a contrast defect that survived two days *\"because the gate that\nswept the page never opened the view\"*. It composes URLs for a reader and stops\nthere.\n\nTHE AGENT CANNOT OPEN THEM. Measured this session, all three routes:\n`curl` -> 403 CONNECT tu", "file": "beans/defs/folio-assistant-7s52--staging-review-hands-a-reader-urls-the-agent-canno.md" }, { "id": "folio-assistant-7sf1", "title": "Move MEMORY.md into the kg graph, with correct skill/task pairings", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-8jt6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T00:04:29Z", "updatedAt": "2026-10-09T19:25:00Z", "preview": "Queued 2026-09-19 while the `skill-not-a-document` split was in flight; not\nstarted. Requested as: *\"move MEMORY.md into kg and make sure in right skill /\ntask pairings\"*.\n\n## What is there now\n\n`.claude/agents/` defines three subagents carrying `memory: project`, each with\na `MEMORY.md` under `.claude/agent-memory//`:\n\n- `platform-boundary-guard` — folio-vs-platform boundary, adapter vs pr", "file": "beans/defs/folio-assistant-7sf1--move-memorymd-into-the-kg-graph-with-correct-skill.md" }, { "id": "folio-assistant-7te5", "title": "who-iris's 12 generated pages are counted as authored — gen-iris-pages marks nothing it writes", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-0lmb", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T06:25:09Z", "updatedAt": "2026-10-07T11:50:39Z", "preview": "Found by running bean `06e3`'s who-iris exercise end to end, 2026-10-04 (PR #2049).\n\n## The measurement\n\n`check:docs-populated` reports for who-iris:\n\n> `✓ who-iris who-iris/docs/kg-to-portal.html — 1491 words of prose`\n> ` 14 authored, 0 generated · who-iris/docs, who-iris/site`\n\nThe true split of those 14 pages is **2 authored, 12 generated**:\n\n| writer | pages |\n|---|---|\n| `who-iris/scrip", "file": "beans/defs/folio-assistant-7te5--who-iriss-12-generated-pages-are-counted-as-author.md" }, { "id": "folio-assistant-7uao", "title": "27 QA witnesses record a stale auditor scriptHash while claiming freshness: fresh", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T02:43:27Z", "updatedAt": "2026-10-09T14:26:00Z", "preview": "Found 2026-10-03 while regenerating after a merge, not by looking for it: a\n`docs:pages` run rewrote 27 witness files, and the ONLY change in each was the\n`scriptHash` of the auditor that produced it.\n\n## The measurement\n\n`auditorHash` is `sha256` of one file and nothing else —\n`cat-harness/scripts/kg-audit.ts:2380`:\n\n```ts\nconst auditorHash = sha256(readFileSync(join(AUDITOR_ROOT, \"scripts\", \"kg-", "file": "beans/defs/folio-assistant-7uao--27-qa-witnesses-record-a-stale-auditor-scripthash.md" }, { "id": "folio-assistant-7w1a", "title": "METHODOLOGIES TABLE MISRENDERS: a truncated cell cuts a code span open and the whole table prints as text", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T18:58:58Z", "updatedAt": "2026-09-24T18:58:58Z", "preview": "Owner 2026-09-24, with a screenshot: \"bean: table misrenders https://litlfred.github.io/folio-assistant/methodologies/ — need QA audit report\".\n\n## Cause\ngen-methodologies-viz.ts short() cuts applies-when at 150 chars. The madr row is cut inside a code span, (see `kepner-tregoe…, and the unmatched backtick pairs with a later one on the line, swallowing the cell pipes. kramdown then renders the who", "file": "beans/defs/folio-assistant-7w1a--methodologies-table-misrenders-a-truncated-cell-cu.md" }, { "id": "folio-assistant-7wgs", "title": "Record when a materialized copy is last READ, so cache:index can rank eviction by use", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T20:03:49Z", "updatedAt": "2026-10-09T19:15:00Z", "preview": "Follow-up from bean `54rk`. `cache:index` has a last-read column and it says `not recorded` for all 276 copies, because nothing records reads.\n\nOwner, 2026-09-23, choosing among \"say not recorded\", \"add a lastReadAt field\" and \"use the file's access time\": **say not recorded**, and open this bean for recording reads later.\n\n**Why not file access time:** it is disabled on many mounts (noatime) and ", "file": "beans/defs/folio-assistant-7wgs--record-when-a-materialized-copy-is-last-read-so-ca.md" }, { "id": "folio-assistant-7wou", "title": "TERMINOLOGY / mapping: check a candidate term against an existing terminology — three states, exact AND concept", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-5yhm", "blocking": [], "blockedBy": [], "createdAt": "2026-09-25T04:51:47Z", "updatedAt": "2026-10-07T19:30:00Z", "preview": "Under `5yhm`, and **the live one** after the 2026-09-29 re-scope. The owner's\nsentence, unchanged: *\"extracting exsiting glossary needs compaision to\nexisting termonology/coding.\"*\n\n## What exists, and the one thing that does not\n\n[`vocabulary-authority`](../../cat-harness/skills/kg/kg-core/vocabulary-authority.md)\nsettles WHICH vocabulary owns a fact — SKOS for meaning, DC for resources,\nFHIR for", "file": "beans/defs/folio-assistant-7wou--terminology-mapping-check-a-candidate-term-against.md" }, { "id": "folio-assistant-7x5n", "title": "SEPARATION ARC: one workplan consolidating five stalled sessions' GOAL 1 work (S0–S8)", "status": "in-progress", "type": "epic", "priority": "high", "parent": "folio-assistant-vuip", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T08:14:10Z", "updatedAt": "2026-10-09T17:47:24Z", "preview": "Consolidates the handovers of five stalled sessions (iirv, w2gr/9umr, cjvs/vke6, fnx4/eayu, d33q) into one arc. Plan, status bars, gap analysis (G1–G14), migration checklist and delegation rules: cat-harness/docs/proposals/separation-arc-2026-10-01.md. Stories S0–S8 are children; each story points at the existing beans it absorbs and owns no new code of its own.\n\nStrategy adopted from sibling bean", "file": "beans/defs/folio-assistant-7x5n--separation-arc-one-workplan-consolidating-five-sta.md" }, { "id": "folio-assistant-7x7g", "title": "external-schemas visualiser: 8 wireframe findings", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T10:36:14Z", "updatedAt": "2026-10-09T14:36:00Z", "preview": "Findings from the as-is wireframe `cat-harness/docs/wireframes/external-schemas/` (intent.md, as-is.html, checks/), observed at 1280×800 and 390×844. Verbatim from its `## Findings`; a finding tagged → is also covered by that cross-cutting bug.\n\n1. **44 of the 53 operative terms say nothing.** Every DCMI term (22) and every BPMN term (22) reads \"derived from the corpus; what this repository does w", "file": "beans/defs/folio-assistant-7x7g--external-schemas-visualiser-8-wireframe-findings.md" }, { "id": "folio-assistant-7x8o", "title": "TRANSLATIONS: 8 pairs may carry less than their source — but 7 of the original 9 were the extractor's hard wrap, not the translators (see lvk9)", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T08:57:09Z", "updatedAt": "2026-10-11T06:23:24Z", "preview": "Found while deriving `.po` catalogues for #206 (bean `tbdg`). Measured\n2026-09-26.\n\n`derive-po.ts` refuses 18 of 25 (page, locale) pairs because the published\ntranslation does not have the same SHAPE as its source. **9 of those 18 are\nartefacts of the extractor** (beans `6b8u` and `ig4a`) and disappear once the\nthreshold is script-neutral. These are the **other 9**, which are still mismatched\nwith", "file": "beans/defs/folio-assistant-7x8o--translations-9-published-pages-carry-less-than-the.md" }, { "id": "folio-assistant-7xmc", "title": "MERGE ROUND TRIM: write the owner-approved trimmed merge-main procedure into merge-conflict-patterns / prepare-merge", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-d33q", "blocking": [], "blockedBy": [], "createdAt": "2026-10-05T05:11:08Z", "updatedAt": "2026-10-07T17:47:00Z", "preview": "Owner ruling 2026-10-05: 'trim duplicated steps'. Measured 2026-10-04 on PR #1898's merge rounds: one round cost ~45 min — merge:main incl. its regen 5–10 min; skill:register 1–2 min; a standalone regen after merge:main 4–5 min (redundant, same tree); full local gates ~24 min (1429 s, bun test on 3–4 cores; CI shards 4 ways in ~5 min); fresh-checkout regen no-op 4–5 min (duplicates CI's clean chec", "file": "beans/defs/folio-assistant-7xmc--merge-round-trim-write-the-owner-approved-trimmed.md" }, { "id": "folio-assistant-7zz1", "title": "STANDALONE: whole-checkout tests move to a declared test home in the top-level folio-assistant instance", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-iirv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T09:15:40Z", "updatedAt": "2026-10-07T17:47:00Z", "preview": "Owner ruling 2026-10-06 ~08:50Z, relayed by the coordinating session (session_012qoycyCSGidZqW245vXhze) and recorded on 0r7u: \"Top-level instance\". The parent folio-assistant checkout, which holds every layer, gets a DECLARED test home: a directory in the top-level instance's declaration, never a literal path in a runner. Basis: whatever names every layer belongs at the top, so each layer stays st", "file": "beans/defs/folio-assistant-7zz1--standalone-whole-checkout-tests-move-to-a-declared-te.md" }, { "id": "folio-assistant-8107", "title": "LMA systems for SE (2404.04834v4): the SDLC landscape review — which stages folio-assistant covers and which it does not", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-0ipy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T23:15:27Z", "updatedAt": "2026-10-09T17:24:00Z", "preview": "Ingest + critical analysis. 30pp systematic review, He/Treude/Lo. The cross-cutting frame: use its SDLC stage map as the gap checklist against this repo's processes/.\n\n## Closed 2026-10-09\n\nIngested arXiv:2404.04834v4 (\"Large Language Model-Based Multi-Agent Systems for Software Engineering: Literature Review, Vision and the Road Ahead\", 30pp, He, Treude, Lo) into `library/arxiv-2404.04834v4/` wit", "file": "beans/defs/folio-assistant-8107--lma-systems-for-se-240404834v4-the-sdlc-landscape.md" }, { "id": "folio-assistant-81qz", "title": "STALE FOLIO PATH IN PLATFORM: audit-wiring.ts's docblock says it walks content/quantum-observable-universe/", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T15:52:13Z", "updatedAt": "2026-10-04T15:52:33Z", "preview": "Recorded from the qou orphaned-content census, 2026-10-04 (ORPH report; session https://claude.ai/code/session_01NdDGeP1SyShmoUssLuRZ91, issue #2106). The census flagged cat-harness/content/pipeline/audit-wiring.ts as walking one folio's tree. Re-measured: the CODE already resolves the paper through requirePaper(--paper) and folioDir. Only the docblock named qou's path (stale since qou moved conte", "file": "beans/defs/folio-assistant-81qz--stale-folio-path-in-platform-audit-wiringtss-docbl.md" }, { "id": "folio-assistant-81t5", "title": "TOOL 8/13: ingest-extract-structure — document ingestion (15 files, 7 entry points)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-d308", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T04:34:56Z", "updatedAt": "2026-09-26T00:00:00Z", "preview": "Group 8 of 13 in `d308`. **15 files, 7 entry points.**\n\n`pdf-extract`, `pdf-ocr`, `pdf-pages`, `pdf-structure`, `pdf-tables`,\n`split-pdf-by-chapter`, `extract-candidates`, `tabular-records`,\n`archive-contents`, `ingest-document`, `scan-repo-content`,\n`smart-base-transform`, `_pdf_doc_id`, `_pypdf_compat`, `_tech_meta`,\n`pypdf_safe`.\n\n**BPMN:** `ingest-extract-structure · Task_ExtractText · Task_Oc", "file": "beans/defs/folio-assistant-81t5--tool-813-ingest-extract-structure-document-ingesti.md" }, { "id": "folio-assistant-81tw", "title": "BOOTSTRAP-TOOLS STAGED AGAIN: the tools that render and validate bootstrap/ leave cat-harness, toward litlfred/bootstrap-tools", "status": "scrapped", "type": "feature", "priority": "high", "parent": "folio-assistant-vuip", "blocking": [], "blockedBy": [], "createdAt": "2026-09-29T23:13:27Z", "updatedAt": "2026-09-30T08:46:25Z", "preview": "**Owner, 2026-09-29:** *\"we will https://github.com/litlfred/bootstrap-tools use this for the tools that render bootstrap/ but were put in cat-harness. for now create dir bootstrap-tools/ for staging split out into repo. make sure IRIs updated\"* — then, choosing scope: *\"1 2 ... need Zod usage in bootstrap tools as part of validation in rendering pipeline. will also need SEMVER skills in bootstrap", "file": "beans/defs/folio-assistant-81tw--bootstrap-tools-staged-again-the-tools-that-render.md" }, { "id": "folio-assistant-81vy", "title": "DEPLOYMENT AWARENESS: the agent must know its UI surface — gh-pages, local server, MCP or chat-only — because it decides which tools apply", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T17:17:15Z", "updatedAt": "2026-10-09T13:33:00Z", "preview": "Owner, 2026-09-20, verbatim:\n\n> agent should be aware of how ui is deployed (e.g. ghpages , no XSS, etc),\n> local server, w/ or w/o mcp, chat discussion only. part of skills to know\n> which tools\n\n## What this adds to the parent epic\n\n`5a3l` establishes that **topology** (where things live) and **operating\nmode** (what the harness is doing) are orthogonal, and that named scenarios\nare their produ", "file": "beans/defs/folio-assistant-81vy--deployment-awareness-the-agent-must-know-its-ui-su.md" }, { "id": "folio-assistant-82qs", "title": "Locale selector back in the LHS navbar's top row", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-9rq1", "blocking": [], "blockedBy": [], "createdAt": "2026-10-05T14:39:33Z", "updatedAt": "2026-10-06T05:55:27Z", "preview": "Owner, 2026-10-05: 'we lost locale selector in top navbar LHS again'. Cause (measured): docs-ui.css ~5729 hides .fa-lang-mini whenever .fa-nav-icons exists; the icon row took the launcher and the light/dark bulb but never a language control. Fix: a globe beside the bulb in the row's after-hook (opens the launcher on the language view), plus an e2e asserting a visible language control in .side-bar ", "file": "beans/defs/folio-assistant-82qs--locale-selector-back-in-the-lhs-navbars-top-row.md" }, { "id": "folio-assistant-872t", "title": "The Python half's WHEEL BUILD is unchecked — rsi6's lesson applied to only one of the two halves", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T09:53:40Z", "updatedAt": "2026-10-09T11:36:00Z", "preview": "Split out of `1s5s` 2026-09-26, while wiring the Python half into\n`check:published-packages`. Recorded rather than decided, because the answer\ncosts something on every CI run and nobody has been asked.\n\n## The asymmetry\n\n`rsi6`'s finding was that **a package that ships and is not built lands green\nand broken** — measured, not predicted: PR #914 bumped `typescript` 5.9.3 →\n7.0.2, passed ten checks,", "file": "beans/defs/folio-assistant-872t--the-python-halfs-wheel-build-is-unchecked-rsi6s-le.md" }, { "id": "folio-assistant-87mi", "title": "Merge refused: #2043 owed CI not green on its head", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T14:30:41Z", "updatedAt": "2026-10-15T16:49:44Z", "preview": "PR #2043, handed back by the Merge Manager on 2026-10-04 (BACKFILLED at 14:50Z: the hand-back was made as a PR comment and a queue entry only, without this bean. Owner: \"Backfill + enforce\").\n\n## Refused\n- `merge:steward` at 14:30Z: Rule_HeadNotGreen: conflict declared, own CI red.\n- Queue entry (`beans/queue/litlfred--folio-assistant--2043.json`, status `waiting-on-author`): STATUS UNKNOWN on the", "file": "beans/defs/folio-assistant-87mi--merge-refused-2043-owed-ci-not-green-on-its-head.md" }, { "id": "folio-assistant-88mg", "title": "CONSOLIDATION: one voice over bootstrap and cat-harness docs, with skills and docs sharing content rather than restating it", "status": "completed", "type": "feature", "priority": "high", "parent": "folio-assistant-2upx", "blocking": [], "blockedBy": [], "createdAt": "2026-09-21T16:21:35Z", "updatedAt": "2026-10-08T04:53:00Z", "preview": "Owner, 2026-09-21: a thorough analysis of the existing documentation AND skills, consolidated, in ONE voice. All of the documentation eventually; **bootstrap and cat-harness first**.\n\n**Two structural asks:**\n- **Reuse content between `skills/` and `docs/`** so both consolidate, with the docs pointing at the skill for rendering where the skill was written correctly.\n- **Some documentation will nee", "file": "beans/defs/folio-assistant-88mg--consolidation-one-voice-over-bootstrap-and-cat-har.md" }, { "id": "folio-assistant-89cl", "title": "STATE BRANCH P5: adjust skills and processes — content-context-and-state-graphs, directory-conventions, todo-manager, bean-coordination, continual-progress, workflow-state, process-state", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T10:58:10Z", "updatedAt": "2026-10-09T15:08:00Z", "preview": "Proposal section 4.1. AFTER the decisions and Phase 3, not before: a skill describing a branch that does not exist yet is a rule nobody can follow.\n\n## Done when\n- [x] each skill in section 4.1 adjusted\n- [x] workflow-state-in-beans.md section 0 addendum\n- [x] AGENTS.md beans section pointers updated\n\nProposal: cat-harness/docs/proposals/state-branch-2026-10-02.md\n\n## Closed 2026-10-09\n\n- Branch: ", "file": "beans/defs/folio-assistant-89cl--state-branch-p5-adjust-skills-and-processes-conten.md" }, { "id": "folio-assistant-89wq", "title": "Four QA checkers are unconditional n/a stubs — 8949 sidecar entries indistinguishable from a correct decline, two of them proof-build-green / proof-no-axiom-growth", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-27T10:37:00Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "Measured 2026-09-27 against this branch's `cat-harness/`, and cross-checked\nagainst qou's committed sidecars at platform pin `37a36ce`.\n\n## Four exported checkers whose entire body is `return { result: \"n/a\", hits: [] }`\n\nFound by parsing every `export function check*` in\n`cat-harness/content/pipeline/qa-checkers*.ts` and keeping those with exactly\none `return` statement, which returns `n/a`:\n\n| c", "file": "beans/defs/folio-assistant-89wq--four-qa-checkers-are-unconditional-na-stubs-8949-s.md" }, { "id": "folio-assistant-8ao5", "title": "Re-seed FHIR AST caches for smart-trust and smart-base after main merge", "status": "scrapped", "type": "task", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T16:04:19Z", "updatedAt": "2026-10-02T16:45:00Z", "preview": "Re-seed the FHIR AST cache orphan branches on litlfred/smart-trust (fhir-ast/smart.who.int.trust) and litlfred/smart-base (fhir-ast/smart.who.int.base) from the current main. The previous seeds were from an earlier commit and may be stale after main merge.\n\n## Done when\n- [x] smart-trust re-seeded — `fhir-ast/smart.who.int.trust` @ `2bc91d0`, 678 resources, 671 edges, 677 fsh-index entries, source", "file": "beans/defs/folio-assistant-8ao5--re-seed-fhir-ast-caches-for-smart-trust-and-smart-base.md" }, { "id": "folio-assistant-8c6v", "title": "MERGE PATTERNS GAP: the 17 generated docs/*.md pages have no declared merge-conflict pattern, so merge:main refuses them although their own front matter says do-not-hand-edit", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-d33q", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T08:55:15Z", "updatedAt": "2026-10-03T11:20:40Z", "preview": "Found 2026-10-03 while measuring PR #1888's merge with `main` for session\n`folio-assistant-a8`: of its 55 conflicts, 53 classified to a declared pattern\nand **2 refused**, and both were the same page in its two forms —\n`cat-harness/content/docs/publication-workflow/every-workflow-in-the-repo.md`\n(authored source) and `cat-harness/docs/process/publication-workflow.md` (its generated\nmirror). The mi", "file": "beans/defs/folio-assistant-8c6v--merge-patterns-gap-the-17-generated-docsmd-pages-h.md" }, { "id": "folio-assistant-8ez4", "title": "STATE BRANCH P0: agree ONE storage field and ONE branch-store library with arc 3fva before either lands", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T10:58:09Z", "updatedAt": "2026-10-06T06:26:35Z", "preview": "Arc 3fva proposes ContentDirectory.storage {branch, keyedBy: commit} for qa-reports. State needs keyedBy: tip (one living tree, splice-write, never -f). One field, one library — a second field over the same question is two answers free to disagree.\n\n## Done when\n- [x] 3fva owner session consulted (PR #1764, comment 2026-10-02)\n- [ ] field shape agreed: 3fva's {branch, keyedBy} widened to commit|ti", "file": "beans/defs/folio-assistant-8ez4--state-branch-p0-agree-one-storage-field-and-one-br.md" }, { "id": "folio-assistant-8fq9", "title": "Placement PR7: tests regroup by concern group, retargeted to cat-harness-tools/", "status": "todo", "type": "task", "priority": "", "parent": "folio-assistant-iirv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T06:58:01Z", "updatedAt": "2026-10-16T06:55:00Z", "preview": "Placement proposal §2 (owner rulings 2026-09-30 §6; process: review → resolve issues → staged PRs). Parent of PR0/PR1 is `9umr`; PR2–PR9 sit under the separation epic because D4 (owner, 2026-10-01, option 3) interleaves them with the split stages into one ordered sequence. Every PR: references move with the subject; generated trees regenerated, never hand-edited; `beans/**` not rewritten; kg-qa si", "file": "beans/defs/folio-assistant-8fq9--placement-pr7-tests-regroup-by-concern-group-retar.md" }, { "id": "folio-assistant-8i57", "title": "Document ingestion page has no theme section: Process_IngestTheme is called but documented nowhere", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T17:22:04Z", "updatedAt": "2026-09-24T18:01:26Z", "preview": "## Why\n`document-ingestion.bpmn` calls `Process_IngestTheme` (`ingest-theme.bpmn`) behind `Gateway_ThemeSource`, and its links pointed at `#ingest-the-theme` and `#a-theme-source` — neither anchor exists. `content/docs/document-ingestion/document-ingestion.ts` has sections for extract / derive / build-KG / gate but none for the theme. Once links are derived from `asset.source` (B5-fix), the theme ", "file": "beans/defs/folio-assistant-8i57--document-ingestion-page-has-no-theme-section-proce.md" }, { "id": "folio-assistant-8iqc", "title": "Library QA after ingestion: titles, bibliographic metadata, placeholder blocks and summary backlog are never judged", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-slw1", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T16:39:43Z", "updatedAt": "2026-10-07T17:24:00Z", "preview": "Issue [#1794](https://github.com/litlfred/folio-assistant/issues/1794).\n\nA library entry is never judged after ingestion or the KG build. Measured on main (59 entries): 24 titles equal the slug, at least 5 are implausible (who-pub-tps-931 is 'Abies', OCR noise off a scanned cover), 0 of 59 carry author or year, and who-pub-tps-931 shows 121 'Page N (no content carried)' blocks over 25,714 OCR word", "file": "beans/defs/folio-assistant-8iqc--library-qa-after-ingestion-titles-bibliographic-me.md" }, { "id": "folio-assistant-8iqt", "title": "recordUntainted writes agent verdicts into the DERIVED block-qa file, not the attestation store", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T19:22:21Z", "updatedAt": "2026-10-06T06:09:21Z", "preview": "Found by the d6bw agent, 2026-10-01.\n\n`recordUntainted` (the untainted-verification writer) still writes an agent verdict into the derived block-qa report. Since 8wj1/2gst, judgements live in the declared `attestations` graph (`test/attestations/`, `qa-attestations/v1`). The verdict reaches the store only when a later writer runs and auto-moves it (owner ruling 2). So nothing is lost today. But on", "file": "beans/defs/folio-assistant-8iqt--recorduntainted-writes-agent-verdicts-into-the-der.md" }, { "id": "folio-assistant-8j9e", "title": "A merge silently drops a gitignored-but-tracked file: green locally, red in CI (#2000)", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-nok9", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T13:33:34Z", "updatedAt": "2026-10-08T01:15:00Z", "preview": "Related: `blgm` (the merge pipeline tools: `merge-base.ts` is where the check belongs). Filed under epic `nok9` because a bean can only have an epic, milestone or feature as its parent.\n\nA merge can silently delete a file that `main` tracks but `.gitignore` covers, and nothing catches it until CI fails on a fresh checkout. Every local check stays green, because regen rewrites the file on disk and ", "file": "beans/defs/folio-assistant-8j9e--a-merge-silently-drops-a-gitignored-but-tracked-fi.md" }, { "id": "folio-assistant-8jt6", "title": "MEMORY & TODOS: notes an agent or a person carries, attached to the graph", "status": "in-progress", "type": "epic", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T11:43:44Z", "updatedAt": "2026-10-05T04:57:14Z", "preview": "Notes that a participant carries, attached to a node of the graph.\n\n`h32d` is the schema both halves share; `5oai` is the relation that lets one\nattach to a content block and move between them. The 2x2 that governs this is\nmemory-vs-workflow and human-vs-agent: todos are human memory, `MEMORY.md` is\nagent memory, beans are agent workflow, and human workflow is deliberately\nempty.\n\n`4kiw` and `0j8h", "file": "beans/defs/folio-assistant-8jt6--memory-todos-notes-an-agent-or-a-person-carries-at.md" }, { "id": "folio-assistant-8jur", "title": "FHIR / IG kinds have no viewer: IG AST, OpenAPI, external schemas, L1 extraction, themes", "status": "in-progress", "type": "feature", "priority": "normal", "parent": "folio-assistant-46ob", "blocking": [], "blockedBy": [], "createdAt": "2026-10-15T16:53:28Z", "updatedAt": "2026-10-22T16:50:00Z", "preview": "## Parent\n\nfolio-assistant-46ob\n\nMeasured live: IG AST ×1 (WHO SMART Trust), Openapi ×1 (WHO SMART Trust), External schemas ×1 (SMART Base), L1 extraction ×1 (SMART Base), Themes ×2 (SMART Base, WHO IRIS).\n\nThe fhir-harness `ig-ast-delta` skill already renders IG ASTs; cat-harness has an `external-schemas` visualiser (check why SMART Base's directory is not covered). Coordinate with the fhir-harne", "file": "beans/defs/folio-assistant-8jur--fhir-ig-kinds-have-no-viewer-ig-ast-openapi-extern.md" }, { "id": "folio-assistant-8lcl", "title": "Separation stage 1b: Zod moves to cat-harness-tools; generated JSON Schema stays in cat-harness", "status": "todo", "type": "task", "priority": "normal", "parent": "folio-assistant-iirv", "blocking": [ "folio-assistant-y9r6" ], "blockedBy": [ "folio-assistant-f8wp" ], "createdAt": "2026-10-01T06:58:01Z", "updatedAt": "2026-10-16T06:55:00Z", "preview": "Stage 1b of the split plan: the 205 Zod `schemas/*.ts` move to `cat-harness-tools/schemas/`; `cat-harness/schemas/` gets **generated** JSON Schema (one per declared node kind, `$id` under `iriBase`), as bootstrap does (`xsqm` ruling 1). `bunfig.toml` preload follows.\n\nOrder: D4 option 3 (owner, 2026-10-01) — after placement PR8, which regroups the same `schemas/` tree and moves the 28 above-harnes", "file": "beans/defs/folio-assistant-8lcl--separation-stage-1b-zod-moves-to-cat-harness-tools.md" }, { "id": "folio-assistant-8mlt", "title": "CREDENTIALS (iv): rename BOOTSTRAP_PAGES_TOKEN -> PUBLISH_SITE_BOOTSTRAP (bootstrap-tools) and MERGE_MAIN_TOKEN -> PUSH_BRANCH_TRIGGERING_CI (merge-main.yml)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T06:31:23Z", "updatedAt": "2026-10-23T08:15:00Z", "preview": "From the proposal cat-harness/docs/proposals/credentials-needs-and-supply.md §4.3 (D2=A). Edit the secret references and the error and help messages in bootstrap-tools/.github/workflows/publish-bootstrap.yml and .github/workflows/merge-main.yml. With the App (D1=A), add the create-github-app-token mint step and expose the job variable under the capability name. Both secrets are unset, so the renam", "file": "beans/defs/folio-assistant-8mlt--credentials-iv-rename-bootstrap-pages-token-publis.md" }, { "id": "folio-assistant-8npa", "title": "Working-notes graph kind: a folio's derivation notes, scoping docs and ledgers as declared context, not rendered content", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-8jt6", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T16:54:09Z", "updatedAt": "2026-10-09T14:01:00Z", "preview": "Downstream need from litlfred/qou (migration story S-REST-1, qou bean qou-pjfi; analysis docs/audits/2026-10-04-workplan-and-migration/REST_report.md in qou).\n\nqou has 512 non-audit docs/ files: 117 math-heavy derivation notes, cheat sheets, scoping docs and recorded refutations, none named by a block. No existing kind fits:\n- `docs` is renderable and would publish working notes as site pages;\n- `", "file": "beans/defs/folio-assistant-8npa--working-notes-graph-kind-a-folios-derivation-notes.md" }, { "id": "folio-assistant-8pyz", "title": "site.data.* undefined renders as empty with no warning: make undefined Liquid variables in IG-site builds a reported finding", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-09T15:38:22Z", "updatedAt": "2026-10-16T06:40:49Z", "preview": "Owner, 2026-10-09, reviewing the smart-* deployments: *\"is a warning made when referenced variable DNE? (QA reports)\"*. It is not, for `site.data.*`.\n\n## What was measured (2026-10-09, gh-pages of litlfred/smart-immunizations @ 51454a6)\n\n`testing.html` carries four links of the form\n`https://raw.githubusercontent.com///main/testing/docker/logic/.env` — the source\n(`input/pagecontent/testing.md`) r", "file": "beans/defs/folio-assistant-8pyz--sitedata-undefined-renders-as-empty-with-no-warnin.md" }, { "id": "folio-assistant-8pzh", "title": "smart-base: extract smart-kg L1 instance documents (publication, section, recommendation) from ingested guideline PDFs", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-qvxh", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T06:16:07Z", "updatedAt": "2026-10-21T21:27:30Z", "preview": "Owner, 2026-10-02 (#1767), option 2 of the smart-kg fold-in: feed smart-kg. smart-kg's README says 'Not built yet: no PDF extractor, so publication and recommendation nodes must currently be authored by hand.' folio-assistant's ingest pipeline already produces sections, outline and page text for each library entry. Build a smart-base tool that turns one ingested guideline (e.g. who-iris/library/97", "file": "beans/defs/folio-assistant-8pzh--smart-base-extract-smart-kg-l1-instance-documents.md" }, { "id": "folio-assistant-8qyc", "title": "REGEN DUPLICATED WORK + BARRIERS: skill:register:check re-runs sub-checks regen also asks; un-barrier read-only checks; narrow inputs", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-xpcu", "blocking": [], "blockedBy": [], "createdAt": "2026-10-05T05:24:07Z", "updatedAt": "2026-10-07T17:47:00Z", "preview": "## Why\nOwner goal 2026-10-05: minimise regen time while staying sound (never a false green). Measured (stream B, main, --jobs 1 --no-cache): 117 pairs, 637 s serial. skill:register:check (107 s, barrier) re-runs ~9 checks regen also asks separately; kg:audit:check (80 s) may be a subset of kg:audit:all:check (198 s); several none/barrier pairs (check:published-instance-exports 49 s, subgraph:jsonl", "file": "beans/defs/folio-assistant-8qyc--regen-duplicated-work-barriers-skillregistercheck.md" }, { "id": "folio-assistant-8rff", "title": "MERGE PATTERNS GAP: three generated families have no declared merge-conflict pattern, so merge:main refuses them and merge:overlap counts them as authored", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-d33q", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T17:46:34Z", "updatedAt": "2026-10-03T00:30:58Z", "preview": "Found by merge:overlap's first live run (bean blgm, PR #1895, 2026-10-02): 32 open PRs; 109 pairs overlapped on authored paths, and 19 of those overlapped ONLY on these three generated families. None is named in cat-harness/scripts/merge-conflict-patterns.ts PATTERNS or in .gitattributes, so they classify as authored.\n\n| family | writer | check in CI | overlap count (pair-path hits) |\n|---|---|---", "file": "beans/defs/folio-assistant-8rff--merge-patterns-gap-three-generated-families-have-n.md" }, { "id": "folio-assistant-8unf", "title": "SESSION LOGS ARE NOT EPICS: todo-manager and session-intent tell every session to mint a 'Session:' milestone", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T15:09:14Z", "updatedAt": "2026-10-04T15:30:37Z", "preview": "Evidence: qou work-plan analysis 2026-10-04 (session https://claude.ai/code/session_01NdDGeP1SyShmoUssLuRZ91). In qou, 292 of 353 epic-typed beans are 'Session:' logs. Cause: todo-manager Core Directive 1 and session-intent step 4b both say `beans create \"Session: \" --type milestone`, and the Status Display sample shows `[epic-123] Session: `. The same skill's §WHICH parent sa", "file": "beans/defs/folio-assistant-8unf--session-logs-are-not-epics-todo-manager-and-sessio.md" }, { "id": "folio-assistant-8v0y", "title": "A bean that RESTATES a live skill's contract is unchecked — kn0t drifted in four places and one turned a measurement into an impression", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T21:07:00Z", "updatedAt": "2026-09-24T05:54:13Z", "preview": "Issue [#1190](https://github.com/litlfred/folio-assistant/issues/1190).\n\nA bean that carries a **skill's contract** instead of pointing at it. `AGENTS.md`'s\nbanner already states the rule — *where a skill and a copy disagree, the skill wins\nand the copy is wrong* — and nothing could read it.\n\n## The defect, already paid for\n\n`kn0t` held a second copy of\n[`fhir-harness/skills/fhir-ig-base/ig-publis", "file": "beans/defs/folio-assistant-8v0y--a-bean-that-restates-a-live-skills-contract-is-unc.md" }, { "id": "folio-assistant-8wj1", "title": "QA READERS F4: block and translation read-modify-write writers stop dropping agent verdicts when the prior is absent — the block-qa D2 split", "status": "completed", "type": "task", "priority": "critical", "parent": "folio-assistant-3fva", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T08:47:13Z", "updatedAt": "2026-10-23T13:35:00Z", "preview": "Arc `3fva`, from reader audit `gxvk` (`cat-harness/docs/proposals/qa-readers-audit-2026-10-01.md`, §5.2 family F4). Refines `oqe3` (`translation:block-qa:check`) and `2ae2` (`qa-agent-write`). As of 2026-10-06 this is not blocked on `16ei`: it completed (qa-store landed; CI publishes `main/` and `pr/` entries and the prune runs on schedule), so the block is withdrawn.\n\n**CRITICAL (C11), and it mus", "file": "beans/defs/folio-assistant-8wj1--qa-readers-f4-block-and-translation-read-modify-wr.md" }, { "id": "folio-assistant-8xtj", "title": "INSTANCE NAMES: folio-assist-core published every core term to a path no term names — fixed, plus the 247-reference residue a sweep must not touch", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T19:01:42Z", "updatedAt": "2026-10-08T01:15:00Z", "preview": "Owner, 2026-09-20 (session_017PqeiS4JYySSWGAYLedmus): *\"fix stub name mismatxhes.\"*\n\n**The load-bearing one is fixed; this bean is the residue**, which needs\nper-site judgement rather than a sweep.\n\n## Fixed already — it was breaking a published artefact\n\n`CORE_NS` is `…/folio-assistant-core/ns#`, so every `fac:` term names that\nstem. Both `docs-site.yml` and `feature-staging.yml` published that l", "file": "beans/defs/folio-assistant-8xtj--instance-names-folio-assist-core-published-every-c.md" }, { "id": "folio-assistant-8xx6", "title": "BPMN re-render for translated labels", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-18T15:07:07Z", "updatedAt": "2026-10-07T17:18:00Z", "preview": "The \"next\" recorded on bean `t8g3`. Re-render BPMN diagrams so translated\nlabels appear, handling text overflow.\n\n**Now unblocked.** Its own input list was partly broken: `schemas/translation-tools.ts`\nlisted `processes/publication-workflow.bpmn` in the `dak` entry's\n`bpmnDiagrams`, and **that file has never existed** — `docs/publication-workflow.md`\nis a PAGE embedding three diagrams. An unresolv", "file": "beans/defs/folio-assistant-8xx6--bpmn-re-render-for-translated-labels.md" }, { "id": "folio-assistant-8zsb", "title": "bun test reads the CHECKOUT PATH: 2 tests fail in any worktree not named folio-assistant", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T17:53:09Z", "updatedAt": "2026-09-30T19:49:03Z", "preview": "## What was measured\n\n`bun test` on `origin/main` `da684477383`, run four times, varying **only the\ndirectory the checkout sits in**. Same commit, same `node_modules`, same\ncommand.\n\n| checkout path | basename | parent | result |\n|---|---|---|---|\n| `/home/user/folio-assistant` | `folio-assistant` | `/home/user` | pass |\n| `/tmp/probe-main/folio-assistant` | `folio-assistant` | `/tmp/probe-main` |", "file": "beans/defs/folio-assistant-8zsb--bun-test-reads-the-checkout-path-2-tests-fail-in-a.md" }, { "id": "folio-assistant-94zs", "title": "REGEN --changed: ask only the pairs a merge touched, and narrow the fixpoint's later passes", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-xpcu", "blocking": [], "blockedBy": [], "createdAt": "2026-10-05T05:12:20Z", "updatedAt": "2026-10-07T17:47:00Z", "preview": "Owner, 2026-10-05: \"maximize efficiency, get regen time as minimal as possible\".\n\n## What\n1. `bun run cat regen --changed `: changed paths = `git diff --name-only ...HEAD` plus the working tree. Ask only pairs whose declared inputs/outputs (globs) or script import closure intersect them. Undeclared and `{tracked}` pairs always run. `--explain` names each skip. Default stays the full ru", "file": "beans/defs/folio-assistant-94zs--regen-changed-ask-only-the-pairs-a-merge-touched-a.md" }, { "id": "folio-assistant-95ir", "title": "Library scanners FILTER OUT a declared-but-absent directory, so a partial checkout reports a clean pass", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T16:29:50Z", "updatedAt": "2026-09-23T21:40:48Z", "preview": "Found by review 2026-09-20, on this branch's own fan-out work (`a02m`).\n\n`check-tabular-stubs.ts:137`, `check-l1-complete.ts:756` and `:936` each end\ntheir directory resolution with a `.filter(existsSync)`. So a library the\ndeclaration names and the tree does not carry is not a finding — it is not\neven a line of output. The check then passes over the libraries that DO exist\nand reports a clean run", "file": "beans/defs/folio-assistant-95ir--library-scanners-filter-out-a-declared-but-absent.md" }, { "id": "folio-assistant-95sk", "title": "FOLIO AS CONVENTION (R24, R27–R29): the folio on any harness, materialised assets and reader documents in folio/, cross-library references — wait on the split", "status": "todo", "type": "task", "priority": "", "parent": "folio-assistant-vke6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T17:59:01Z", "updatedAt": "2026-09-23T17:59:01Z", "preview": "Carried over from `j2if` on 2026-09-23, on the owner's choice \"Move to split, close\". The source of truth for the wording is `cat-harness/docs/architecture/folio-board-requirements.md` §\"The folio is a convention, not a page — R24 to R29\". Quoted here so this bean can be read on its own:\n\n| | requirement | state 2026-09-23 |\n|---|---|---|\n| **R24** | the folio visualisation is `cat-harness`'s, and", "file": "beans/defs/folio-assistant-95sk--folio-as-convention-r24-r27r29-the-folio-on-any-ha.md" }, { "id": "folio-assistant-96y4", "title": "smart-immunizations has no BPMN: IMMZ.A-I exist only as SVG images, and its personas only as HTML", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-10T14:06:52Z", "updatedAt": "2026-10-21T12:40:00Z", "preview": "Measured 2026-10-10 on litlfred/smart-immunizations main (86898e6): its nine processes IMMZ.A–I exist only as SVG images (`input/images/immz-*-business-process.svg`) and a pagecontent table. `input/business-processes/` is empty and there are no `.bpmn` files. The personas are HTML tables only; `input/fsh/actors/` is empty.\n\nSo the recommendation-scenario extraction can reference a process only by ", "file": "beans/defs/folio-assistant-96y4--smart-immunizations-has-no-bpmn-immza-i-exist-only.md" }, { "id": "folio-assistant-9791", "title": "QA SUMMARY: 'passing' counts only script-checkable criteria — agent-judged criteria never run are not said (heat map reads it as QA passed)", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-q4jm", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T18:10:15Z", "updatedAt": "2026-10-07T18:03:00Z", "preview": "Found while doing tw61, 2026-09-23. On a freshly scaffolded document folio, a sweep records 46 criteria per block, and every recorded one passes or is n/a. `publish-block-qa` then says **\"passing\"**. But about 15 criteria that need an agent's judgement were never run: voice traps, exposition clarity, adversarial review. They appear in the sweep output as `[needs-agent]` and nowhere in the verdict ", "file": "beans/defs/folio-assistant-9791--qa-summary-passing-counts-only-script-checkable-cr.md" }, { "id": "folio-assistant-99vu", "title": "site.data.fhir from the FHIR AST, not sushi-config: one source for the site's IG variables and its artefact pages", "status": "scrapped", "type": "feature", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-09T15:38:22Z", "updatedAt": "2026-10-09T15:43:21Z", "preview": "Owner, 2026-10-09: *\"site.data.fhir should come fhir AST or so...\"*.\n\n## Today (measured 2026-10-09, litlfred/fhir-harness HEAD)\n\n`fhir-harness/scripts/ig-site-data.ts` (bean `bamf`) fills `site.data.fhir` from\n`sushi-config.yaml`, else `fhir-artifact-index/index.json`, and writes only\n`ig.{id,url,name,title,version,status,publisher,fhirVersion}`, `packageId` and\n`canonical`; everything else is li", "file": "beans/defs/folio-assistant-99vu--sitedatafhir-from-the-fhir-ast-not-sushi-config-on.md" }, { "id": "folio-assistant-9c7h", "title": "Move fsh-guts/ to its own special branch cat/cat-harness/fsh-guts (separation prerequisite); retarget its tools", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-7x5n", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T21:20:17Z", "updatedAt": "2026-10-04T12:52:02Z", "preview": "Owner, 2026-10-02, verbatim: \"bean: as pre-requistie as part of separation, fsh-guts content gets its own named branch, cat/cat-harness/fsh-guts and the contents of fsh-guts/ dir goes there. will need to update the tools associated to this, like with other special named branches.\"\n\nBranch name updated 2026-10-02 per the owner's naming ruling, `cat//` (note on `fs43`). The owner aske", "file": "beans/defs/folio-assistant-9c7h--move-fsh-guts-to-its-own-special-branch-cat-fsh-gu.md" }, { "id": "folio-assistant-9cc0", "title": "ROOT TYPECHECK IS RED ON MAIN: tsconfig sweeps a separately-published sub-package whose vitest dep is not installed at the root", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T11:06:36Z", "updatedAt": "2026-09-26T12:54:56Z", "preview": "## What is failing\n\n`bunx tsc --noEmit` exits **2** on `main`:\n\n cat-harness/schemas/block-qa-schema/tests/parity.test.ts(26,40):\n error TS2307: Cannot find module 'vitest' or its corresponding type declarations.\n\n**Measured on pristine `origin/main` at `a7b5f021498c786a90198196e18eb3d6c27a61ce`\nin a clean detached worktree with `bun install --frozen-lockfile` — not on a\nbranch, not in a dir", "file": "beans/defs/folio-assistant-9cc0--root-typecheck-is-red-on-main-tsconfig-sweeps-a-se.md" }, { "id": "folio-assistant-9f05", "title": "Merge refused: #2063 not signed (merge:guard checks 3 and 4)", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-d33q", "blocking": [ "folio-assistant-0qjq" ], "blockedBy": [], "createdAt": "2026-10-04T14:30:41Z", "updatedAt": "2026-10-07T14:55:47Z", "preview": "PR #2063, handed back by the Merge Manager on 2026-10-04 (BACKFILLED at 14:50Z: the hand-back was made as a PR comment and a queue entry only, without this bean. Owner: \"Backfill + enforce\").\n\n## Refused\n- `merge:steward` at 14:30Z: Rule_NotReady: conflict clean, own CI green.\n- Queue entry (`beans/queue/litlfred--folio-assistant--2063.json`, status `active`): STATUS UNKNOWN on the 13:40Z all-PR r", "file": "beans/defs/folio-assistant-9f05--merge-refused-2063-not-signed-mergeguard-checks-3.md" }, { "id": "folio-assistant-9fdi", "title": "Task_Review needs theme-ui-review, which NO role carries — and the ruling to re-lane it contradicts the diagram's own no-call reason", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T16:30:53Z", "updatedAt": "2026-09-23T11:30:00Z", "preview": "Found 2026-09-23 by stream 3/3 of the #956 consolidation, while applying the\nowner's ruling on the `role-carries-activity-skill` findings. **Five of the six\nwere closed by that ruling. This is the sixth, and it is stuck for a reason the\nquestion did not know about.**\n\n## What was fixed, and what this is\n\n`ingestion-agent` now carries `theme-art-intake`, which cleared five findings in\none line — th", "file": "beans/defs/folio-assistant-9fdi--task-review-needs-theme-ui-review-which-no-role-ca.md" }, { "id": "folio-assistant-9gkj", "title": "Merge refused: #2094 owed CI not green on its head", "status": "completed", "type": "bug", "priority": "", "parent": "folio-assistant-1xhc", "blocking": [ "folio-assistant-wekz" ], "blockedBy": [], "createdAt": "2026-10-04T14:30:43Z", "updatedAt": "2026-10-07T14:55:47Z", "preview": "PR #2094, handed back by the Merge Manager on 2026-10-04 (BACKFILLED at 14:50Z: the hand-back was made as a PR comment and a queue entry only, without this bean. Owner: \"Backfill + enforce\").\n\n## Refused\n- `merge:steward` at 14:30Z: Rule_HeadNotGreen: conflict declared, own CI missing-required.\n- Queue entry (`beans/queue/litlfred--folio-assistant--2094.json`, status `active`): ACK: Owner-approved", "file": "beans/defs/folio-assistant-9gkj--merge-refused-2094-owed-ci-not-green-on-its-head.md" }, { "id": "folio-assistant-9gyz", "title": "QA-review schemas: Finding, Decision and AuditNote are three entities, not one", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-1swy", "blocking": [], "blockedBy": [], "createdAt": "2026-09-18T20:09:07Z", "updatedAt": "2026-10-07T17:18:00Z", "preview": "## Progress — 2026-09-18\n\nStep 1 landed on `claude/festive-galileo-s7ibx0` ([PR #280](https://github.com/litlfred/folio-assistant/pull/280)), two commits.\n\n**`schemas/qa-review.ts`** — `Finding`, `Decision`, `AuditNote` as three\nentities. The two severity vocabularies are kept as orthogonal axes:\n`critical | major | minor` is what kind of breakage (a machine can assign it),\n`blocking | suggestion ", "file": "beans/defs/folio-assistant-9gyz--qa-review-schemas-finding-decision-and-auditnote-a.md" }, { "id": "folio-assistant-9hfi", "title": "smart-immunizations codings.html: CodeSystems and ValueSets show 'not rendered: list-simple-*.xhtml'", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T05:56:52Z", "updatedAt": "2026-10-06T18:30:00Z", "preview": "Owner, 2026-10-06: on https://litlfred.github.io/smart-immunizations/codings.html both lists render as placeholders:\n\n CodeSystems ✎ 📣\n ⟦not rendered: list-simple-codesystems.xhtml⟧ — the IG Publisher renders it from a source this build does not hold; no renderer was given to this build.\n ValueSets ✎ 📣\n ⟦not rendered: list-simple-valuesets.xhtml⟧ — (same)\n\nThe owner asked \"what to do", "file": "beans/defs/folio-assistant-9hfi--smart-immunizations-codingshtml-codesystems-and-va.md" }, { "id": "folio-assistant-9hlq", "title": "Provision gh-pages before Pages is switched on, in every publishing route (#2417)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-0mpw", "blocking": [], "blockedBy": [], "createdAt": "2026-10-07T12:06:33Z", "updatedAt": "2026-10-08T04:30:00Z", "preview": "Issue #2417. FR-1..FR-4: pages-bootstrap unprovisioned state + --provision; getting-started.bpmn Provision gh-pages task; skills getting-started/repo-conversion/remote-mount name the step. Owner rulings 2026-10-01 and 2026-10-07.\n\n## Completed on landed evidence\nLanded on main in PR #2417 (commit fa491d3d5a0c, \"Provision gh-pages before Pages is switched on, in every publishing route (#2417)\").\n- ", "file": "beans/defs/folio-assistant-9hlq--provision-gh-pages-before-pages-is-switched-on-in.md" }, { "id": "folio-assistant-9hxd", "title": "CREDENTIALS (iii): credentials/ declaration (needs, profile, registry) for bootstrap-tools and folio-assistant", "status": "todo", "type": "task", "priority": "", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T06:31:22Z", "updatedAt": "2026-10-16T10:10:00Z", "preview": "From the proposal cat-harness/docs/proposals/credentials-needs-and-supply.md §2.3–2.4 (D3=A). Register three graph kinds: credential-needs (context), credential-profile (context), credential-registry (state). Write the needs for every secret in §1.1. The profile is personal-github with the App (D1=A). Declare sign-artefact as a need with no supply yet (D4=A). bootstrap-tools carries its own file, ", "file": "beans/defs/folio-assistant-9hxd--credentials-iii-credentials-declaration-needs-prof.md" }, { "id": "folio-assistant-9io2", "title": "lake-cache family to cat/folio-assistant-sci/lake-cache/ (declaration, 8 mirrors, folio-repo renames)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T00:51:43Z", "updatedAt": "2026-10-07T19:30:00Z", "preview": "The last special branch still declared under the interim `cat-` scheme. Owner, 2026-10-02: special branches are `cat//`; and (confirmed 2026-10-02, Parcel B brief) folio-assistant-sci is the lake-cache harness. So the family becomes `cat/folio-assistant-sci/lake-cache/-`.\n\nSplit out of 9c7h on 2026-10-03 because it touches 8 CI and script files that run INSIDE folio", "file": "beans/defs/folio-assistant-9io2--lake-cache-family-to-catfolio-assistant-scilake-ca.md" }, { "id": "folio-assistant-9k2i", "title": "CROSS-PAPER synthesis + adversarial pass ON folio-assistant: what transfers, what is refuted, what gap remains", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-0ipy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T23:15:27Z", "updatedAt": "2026-10-09T18:05:00Z", "preview": "Depends on the five per-paper beans. Must keep measured / recommended / claimed apart, as methodologies/merge-queue.md does. Adversarial in both directions: attack the papers' claims, and attack this repo's processes with them.\n\n## Closed 2026-10-09\n\nSynthesized the five foundational agentic SE papers (arXiv:2404.04834v4, arXiv:2507.23348v1, arXiv:2607.00053v1, arXiv:2601.04544v1, arXiv:2402.02172", "file": "beans/defs/folio-assistant-9k2i--cross-paper-synthesis-adversarial-pass-on-folio-as.md" }, { "id": "folio-assistant-9mmu", "title": "GENERIC ORPHAN DETECTION: one detector for an artefact that names a subject that is gone, across graph kinds", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-10-04T15:10:08Z", "updatedAt": "2026-10-09T18:05:00Z", "preview": "Recorded from the qou work-plan analysis, 2026-10-04 (session https://claude.ai/code/session_01NdDGeP1SyShmoUssLuRZ91). Not started: recorded so the gap has an owner. Orphan selectors exist per artefact family (s8nu found four for one question). qou showed orphans in beans (parents gone), branches (work not on main), and block sidecars. Proposed: one kg:audit criterion family over every declared g", "file": "beans/defs/folio-assistant-9mmu--generic-orphan-detection-one-detector-for-an-artef.md" }, { "id": "folio-assistant-9ofm", "title": "STATE BRANCH P3: migrate every reader and writer of beans/, todos/, issue-marks/, health results to the branch", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-fs43", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T10:58:10Z", "updatedAt": "2026-10-07T17:37:00Z", "preview": "Proposal section 4 table, one row per change: .beans.yml, workflow/store.ts WORKFLOW_DIR, claim-bean.ts, beans-fallback.ts, beans-landed.ts, gen-docs-pages.ts, state-visualizer.ts, test/health/run.ts, todos.ts, render-pipeline.ts, issue-marks readers, folio_init templates.\n\n## Done when\n- [ ] every row green against the branch\n\nProposal: cat-harness/docs/proposals/state-branch-2026-10-02.md\n\n_2026", "file": "beans/defs/folio-assistant-9ofm--state-branch-p3-migrate-every-reader-and-writer-of.md" }, { "id": "folio-assistant-9rnf", "title": "DISCOVERY: publishedPairs scans the site root only, so a translated page in a subdirectory is invisible to every catalogue tool", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-27T04:59:40Z", "updatedAt": "2026-09-27T08:07:02Z", "preview": "## The gap\n\n`publishedPairs` in `cat-harness/content/pipeline/derive-po.ts` reads the site\ndirectory with a single `readdirSync`, so it sees pages at the root and nothing\nbelow it. Measured on `main` at `c6960465301`:\n\n| | |\n|---|---|\n| published translated pages | **14** |\n| of those, found by discovery | 13 (top level) |\n| **invisible** | **1** — `guides/agent-onboarding.md`, in ar, es, fr, ru, ", "file": "beans/defs/folio-assistant-9rnf--discovery-publishedpairs-scans-the-site-root-only.md" }, { "id": "folio-assistant-9rq1", "title": "ONE NAVBAR: every navbar capability is a harness feature, drawn by one mechanism on every layout (.side-bar, .fa-nav rail, viewer pages)", "status": "in-progress", "type": "feature", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-10-05T05:37:13Z", "updatedAt": "2026-10-21T21:32:00Z", "preview": "Owner, 2026-10-05, choosing how to fix the missing harness icon row on who-iris pages (#2147): '1... this should be a common navbar funcationliatyt in harness. should be a bean about this'.\n\n## The gap this names\nThe navbar's capabilities are defined per LAYOUT, not per harness. The harness icon row (mountNavIconRow in docs-ui.js) binds only just-the-docs's .side-bar and reads data that only head_", "file": "beans/defs/folio-assistant-9rq1--one-navbar-every-navbar-capability-is-a-harness-fe.md" }, { "id": "folio-assistant-9rrb", "title": "CATALOGUE DISCOVERY: derive-po held a list of five page names, so #1404's four new pages took drift from 1 to 21 unseen", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T14:13:00Z", "updatedAt": "2026-10-07T05:05:00Z", "preview": "Found 2026-09-26, minutes after #1369 merged, by watching what happened next.\n\n#1369 took `translation-drift` from 25 findings to 1. **PR #1404 then merged four\nmore translated pages across five locales with no catalogues, and the gate went to\n21.** The tool that had just catalogued 24 pairs could not see any of them, because\nits CLI held a list:\n\n```ts\nconst pages = [\"accessibility\", \"content-typ", "file": "beans/defs/folio-assistant-9rrb--catalogue-discovery-derive-po-held-a-list-of-five.md" }, { "id": "folio-assistant-9scf", "title": "folio visualiser: 5 wireframe findings", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-4ccr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T10:36:14Z", "updatedAt": "2026-10-09T14:40:00Z", "preview": "Findings from the as-is wireframe `cat-harness/docs/wireframes/folio/` (intent.md, as-is.html, checks/), observed at 1280×800 and 390×844. Verbatim from its `## Findings`; a finding tagged → is also covered by that cross-cutting bug.\n\n1. **Horizontal scroll at phone width.** At 390 px the document is 769 px wide (`scrollWidth` 769). The node table's *declared in*, *links* and *prose* columns start", "file": "beans/defs/folio-assistant-9scf--folio-visualiser-5-wireframe-findings.md" }, { "id": "folio-assistant-9udd", "title": "LSI: cross-document link proposals need a minimum cosine and a hub penalty", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ansc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-29T21:51:12Z", "updatedAt": "2026-09-29T23:23:22Z", "preview": "who-iris corpus: of the 15 strongest LSI cross-document pairs, 4 are real, 3 partial, 8 spurious; WPRO p3 is the nearest cross-document unit for 56 units and HQ p78 for 46 (hubs); median best Handbook->HQ cosine 0.195 — 'nearest' mostly means 'least unlike'. Done when: any cross-document proposal (graph-search --latent, lsi_query across graphs, future uses[] candidates) applies a floor and a hub p", "file": "beans/defs/folio-assistant-9udd--lsi-cross-document-link-proposals-need-a-minimum-c.md" }, { "id": "folio-assistant-9umr", "title": "Concern subgraphs: cat-harness skills live in one declared subgraph per semantic concern (first: KG + library out of folio-core)", "status": "in-progress", "type": "epic", "priority": "normal", "parent": "folio-assistant-vuip", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T08:12:08Z", "updatedAt": "2026-10-21T17:24:00Z", "preview": "Owner 2026-09-29: 'declared sub-graphs of cat-harness based on semantic concern (see sibling work on bootstrap definitions as named subset). dont need a separate facet. built into location.' Ruled 2026-09-30: **eight groups, KG + library move first**.\n\nA Subgraph is bootstrap's term (cat-harness/schemas/graph.ts): 'a named subset of Node Instances: a Declaration's directory entry, with its id, its", "file": "beans/defs/folio-assistant-9umr--concern-subgraphs-cat-harness-skills-live-in-one-d.md" }, { "id": "folio-assistant-9v4m", "title": "TEST INTERFERENCE: a profile-gate test fails in the full suite and passes in isolation on the same commit", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T03:40:20Z", "updatedAt": "2026-10-07T17:37:00Z", "preview": "Measured 2026-09-26 on `985d0bada2a` (branch claude/brave-hypatia-r820sf, diff = two bean .md files only).\n\n## What happened\n\n`bun run cat gates` failed with ONE test failure:\n\n (fail) the sweep's profile gate, end to end > a paper-only criterion is n/a'd in a document folio, under its OWN outcome [5769.02ms]\n\nThe same test, by name, **passes in isolation** — both with the branch's changes appl", "file": "beans/defs/folio-assistant-9v4m--test-interference-a-profile-gate-test-fails-in-the.md" }, { "id": "folio-assistant-9v5a", "title": "QA BACKFILL: per-content-block adversarial QA over the existing corpus, plus the methodology research it rests on", "status": "completed", "type": "epic", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T22:27:08Z", "updatedAt": "2026-10-10T08:00:00Z", "preview": "Owner, 2026-10-02: *\"Split into TWO epics: gates, and QA backfill\"*. This is the\nsecond half. `nok9` keeps the merge-GATE half.\n\n## Why the cut is here and not elsewhere\n\n`nok9` as written mixed two things with different cadences:\n\n- **Gates** block a merge. They are a merge-pipeline change, they run on a diff,\n and they are judged per PR. `nok9`'s children A, B, C and E are these.\n- **Backfill**", "file": "beans/defs/folio-assistant-9v5a--qa-backfill-per-content-block-adversarial-qa-over.md" }, { "id": "folio-assistant-9wb0", "title": "PUBLICATION IS A STATE, NOT A BOOLEAN: every asset carries id + version and sits in draft; formal publication is an undefined process", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-vke6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T21:29:44Z", "updatedAt": "2026-10-09T16:40:00Z", "preview": "Owner ruling, 2026-09-23: *\"all assets get a version and are in 'draft' publication. formal publication process needs to be deinfed/neeeds tools/depends on instance\"*.\n\nCorrects §3.1 as `snjh` shipped it. Not a gap in that work — a different model.\n\n## What §3.1 got wrong, precisely\n\nIt models `publishable?: boolean` with three states, and REFUSES id and version unless `publishable: true`:\n\n re", "file": "beans/defs/folio-assistant-9wb0--publication-is-a-state-not-a-boolean-every-asset-c.md" }, { "id": "folio-assistant-9x01", "title": "available_locales claims a locale with no translated page behind it, and nothing checks the two against each other", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T14:26:44Z", "updatedAt": "2026-09-26T20:14:52Z", "preview": "Found while fixing an e2e fixture (bean `6bhf`, PR #1408), not while looking for\nit — so the measurement below is narrow on purpose and the sweep is part of the\nwork.\n\n## The contradiction\n\n`cat-harness/docs/process/crdm-methodology.md` declares\n\n available_locales: [\"en\",\"fr\"]\n\nand there is **no `cat-harness/docs/fr/crdm-methodology.md`**. It is also absent\nfrom `docs/_data/translations.json`,", "file": "beans/defs/folio-assistant-9x01--available-locales-claims-a-locale-with-no-translat.md" }, { "id": "folio-assistant-9x9r", "title": "check:head-has-run passes on ANY run naming the sha — two unrelated push runs satisfy it while the gate workflow never fired", "status": "completed", "type": "bug", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T19:38:43Z", "updatedAt": "2026-09-25T18:14:49Z", "preview": "Measured 2026-09-24, on PR #1309's head. The script printed a ✓, and the gate\nset had never fired.\n\n## The pass condition is strictly weaker than the defect it detects\n\n`runsForHead` asks the Actions API for every run whose `head_sha` is this\ncommit, and line 191 decides:\n\n```ts\nreturn runs.length > 0 ? { state: \"has-run\", runs } : { state: \"no-run\" };\n```\n\nAny run. Any workflow, any event. The he", "file": "beans/defs/folio-assistant-9x9r--checkhead-has-run-passes-on-any-run-naming-the-sha.md" }, { "id": "folio-assistant-9y9j", "title": "JSON-LD VOICE: the rules every emitted JSON-LD document is written and reviewed against, cited to the held W3C texts", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-scfh", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T16:24:38Z", "updatedAt": "2026-10-01T17:29:04Z", "preview": "Owner, 2026-10-01: 'make sure that these issues are part of the one voice for jsonld'; then chose option 1 — a new voice, with 'jsonld' added as a voice target, owned by cat-harness, each rule cited (check:voices enforces citations).\n\nRules (from jcet and the PROV measurement, each to be cited verbatim to a held source):\n1. object properties are links (@type @id), never literals — measured: 100 ac", "file": "beans/defs/folio-assistant-9y9j--json-ld-voice-the-rules-every-emitted-json-ld-docu.md" }, { "id": "folio-assistant-9zok", "title": "bun run cat gates can NEVER pass translation:catalogue:check — it scrapes a command whose $base is a shell variable defined on an earlier line of the workflow", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T18:56:04Z", "updatedAt": "2026-10-07T04:51:03Z", "preview": "## The defect\n\n`AGENTS.md` makes `bun run cat gates` the STRICT pre-push command, and its whole\nargument for existing is that `gates.ts` derives its list from\n`.github/workflows/code-quality-gates.yml` so it *\"cannot drift from what CI\nactually runs\"*. For one gate the derivation is lossy in a way that makes the\ngate unpassable locally, for everyone, always.\n\n`code-quality-gates.yml:1341-1348` is ", "file": "beans/defs/folio-assistant-9zok--bun-run-gates-can-never-pass-translationcataloguec.md" }, { "id": "folio-assistant-a0s3", "title": "B8 (#1168): string-to-reference sweep — typed refs on ~25 bare-string fields; delete SkillDefinition.roles and front-matter package:", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-tr05", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T21:12:05Z", "updatedAt": "2026-09-30T10:42:13Z", "preview": "Mechanical sweep from the string-to-reference analysis (#1168 comment 5803006038). SkillNameSchema / ProcessIdSchema / TaskRefSchema / BeanIdSchema / BLOCK_KINDS enum / ThemeRefSchema / KgRefSchema on the listed fields; kg-export ActorDef.roles and role toJsonLd skills as @id links; resolve memory roles/agents. Delete SkillDefinition.roles (inverse of RoleDef.skills, 0 instances) and skill front-m", "file": "beans/defs/folio-assistant-a0s3--b8-1168-string-to-reference-sweep-typed-refs-on-25.md" }, { "id": "folio-assistant-a13a", "title": "TERMINOLOGY / identification: say what a term IS, per asset kind — and what is not", "status": "scrapped", "type": "task", "priority": "normal", "parent": "folio-assistant-5yhm", "blocking": [], "blockedBy": [], "createdAt": "2026-09-25T04:51:47Z", "updatedAt": "2026-09-29T22:01:05Z", "preview": "**SCRAPPED 2026-09-29 — superseded on `main` before this bean was ever\ncommitted.** Not deleted: a scrapped bean stops the next agent re-entering a\ndead end, and this one would read as an obvious gap to anybody who had not\nchecked.\n\nWritten 2026-09-25 to ask \"what IS a term, per asset kind, and what is not\",\non the reading that `glossary-page.ts`'s rule was implicit. It is not implicit\nany more. `", "file": "beans/defs/folio-assistant-a13a--terminology-identification-say-what-a-term-is-per.md" }, { "id": "folio-assistant-a1ku", "title": "Generate a paper's leanblueprint layout (blueprint/src) from its manifest, and wire it into the folio blueprint.yml", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T13:39:21Z", "updatedAt": "2026-09-30T15:42:28Z", "preview": "Owner 2026-09-30: after #1591, 'fix the platform (1), then the blueprint wiring (3)'. Parent work: yxob / 6dvy (#1492).\n\n## Why\nThe folio blueprint.yml runs docgen-action with blueprint: true, which needs /lean/blueprint/src/ (web.tex, print.tex, content.tex, macros/, plastex.cfg). Nothing generates it: blueprint-export.ts writes content only, and the one blueprint/src in the tree (cat-harn", "file": "beans/defs/folio-assistant-a1ku--generate-a-papers-leanblueprint-layout-blueprintsr.md" }, { "id": "folio-assistant-a4of", "title": "S1 bookkeeping truth: close merged beans, reconcile fnx4 boxes, fold w2gr 3b into 70lx", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-7x5n", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T08:14:33Z", "updatedAt": "2026-10-06T19:30:00Z", "preview": "G2+G3 of the arc plan.\n\n## Done when\n- [x] y5si, ybp4 completed with PR evidence (#1599, #1687)\n- [x] ejye, 7dek completed (#1758, #1760); ~~ybwt~~ carried on its own bean, see the 2026-10-06 summary. Its one open box is code work, not bookkeeping\n- [x] fnx4 boxes ticked for slices merged in #1721\n- [x] w2gr step 3b list recorded on 70lx so the move happens once\n\n\n## The sweep, measured 2026-10-03", "file": "beans/defs/folio-assistant-a4of--s1-bookkeeping-truth-close-merged-beans-reconcile.md" }, { "id": "folio-assistant-a7t8", "title": "ingest-document's 'run the remaining arms' line double-nests the pdf-images output", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-slw1", "blocking": [], "blockedBy": [], "createdAt": "2026-09-22T20:13:04Z", "updatedAt": "2026-09-30T00:28:36Z", "preview": "Hit while ingesting the SWOT paper, 2026-09-22.\n\n## What happened\n\n`bun run cat ingest ` stages, then prints:\n\n> Next: run the remaining arms with `-o ingest-staging/`, then: ... --promote\n\nFollowing that literally:\n\n```\npython3 cat-harness/scripts/pdf-images.py -o cat-harness/ingest-staging/gurel-tat-2017-swot-analysis \n wrote cat-harness/ingest-staging/gurel-tat-2017-swot-analys", "file": "beans/defs/folio-assistant-a7t8--ingest-documents-run-the-remaining-arms-line-doubl.md" }, { "id": "folio-assistant-a8wy", "title": "VECTOR FIGURE ARM: extract positioned labels from vector-only figures — 69 of 87 caption pages", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-2yyh", "blocking": [], "blockedBy": [], "createdAt": "2026-09-24T18:04:22Z", "updatedAt": "2026-10-06T19:30:00Z", "preview": "Sibling of `m4xy` under the same epic, and the second of its two open arms, and the narrow one: **the labels, and\nnothing above them.** `pdf-images.py` recovers the raster layer; a WHO\nconceptual figure is drawn in path operators and text, so there is no image\nobject and that arm places nothing. `9789240120747-eng` declares six captioned\nfigures and places zero images, and `check-l1-complete` repo", "file": "beans/defs/folio-assistant-a8wy--vector-figure-arm-extract-positioned-labels-from-v.md" }, { "id": "folio-assistant-a98i", "title": "translate-kg-viewer --extract refreshes every .pot but never syncs the .po stubs", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-bzyu", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T08:11:21Z", "updatedAt": "2026-10-07T05:09:00Z", "preview": "MEASURED 2026-09-19 while adding one string to `scripts/kg-viewer-strings.ts`.\n\n`bun run cat translate-kg-viewer --extract` rewrites all five `translations//kg-viewer.pot` files and refreshes each TranslationNode manifest. It does **not** touch the `.po` stubs beside them. So a new msgid leaves every catalogue one entry short, and nothing in the tool says so — it prints `po: 0/39, manifest re", "file": "beans/defs/folio-assistant-a98i--translate-kg-viewer-extract-refreshes-every-pot-bu.md" }, { "id": "folio-assistant-a9tx", "title": "IG PUBLISHER FORK: requirements for an agent working a local experimental fork, and what the AST must carry", "status": "todo", "type": "feature", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [ "folio-assistant-h3tx" ], "blockedBy": [], "createdAt": "2026-09-22T19:07:23Z", "updatedAt": "2026-10-23T20:40:00Z", "preview": "> **Taking this over? Start at §\"HANDOVER — start here (2026-10-01)\" at the end of this bean.** Earlier sections are history; the order list in §\"Prepared for the agent with network\" is superseded.\n\nRequirements for an agent working a **local experimental fork** of the IG\nPublisher, and what the AST it emits must carry.\n\nScope this round: **requirements only, no Java.**\n\n## Why a fork rather than ", "file": "beans/defs/folio-assistant-a9tx--ig-publisher-fork-requirements-for-an-agent-workin.md" }, { "id": "folio-assistant-aazi", "title": "WORKFLOW-STATEMGMT RENDERING: a status dashboard atop the root README, and beans/todos mapped to their BPMNs", "status": "completed", "type": "task", "priority": "", "parent": "folio-assistant-yj32", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T17:11:46Z", "updatedAt": "2026-10-07T19:30:00Z", "preview": "Owner, 2026-09-20, verbatim:\n\n> if a harness is working/active KG repo (not static read only)repo root READM\n> should show at top a current status of all intiated/partially iniated/etc.\n> quick dashboard (#todos, #beans) for each initiated harness. todo on render\n> of active KG, beans and todos mapped to thier bpamns rendered nicely as well\n> for project managemtn / visualizaitons. available a", "file": "beans/defs/folio-assistant-aazi--workflow-statemgmt-rendering-a-status-dashboard-at.md" }, { "id": "folio-assistant-abmq", "title": "MERGE GATE (c): RED FLAG taxonomy, verdict sidecar shape, and the recorded override path", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-nok9", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T16:29:16Z", "updatedAt": "2026-10-22T16:45:00Z", "preview": "Child (c) of the merge-gate epic. Design: `cat-harness/docs/proposals/merge-gate-2026-10-02.md` §6.\n\nA RED FLAG is a finding the reviewer asserts **would block the merge** (amended 2026-10-03; it asserted \"blocks the merge\" until the warn-only ruling below). It reuses the two existing axes in `schemas/qa-review.ts`: `FindingSeverity` (critical | major | minor: what kind of breakage) and `FindingWe", "file": "beans/defs/folio-assistant-abmq--merge-gate-c-red-flag-taxonomy-verdict-sidecar-sha.md" }, { "id": "folio-assistant-afu3", "title": "SEARCH PINNED: the magnifier stays on screen at every scroll position (#1732)", "status": "completed", "type": "feature", "priority": "high", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-01T00:56:04Z", "updatedAt": "2026-10-07T20:30:00Z", "preview": "Owner, 2026-10-01: the search magnifier should ALWAYS be on screen, mid-page included. Issue #1732, follow-up to #1715/#1720. Branch claude/search-pinned.\n\n## Completed on landed evidence\n\n- Implementation landed on `main` in PR #1732 (commit `49828a5686a3`): search magnifier pinned to stay visible on screen across all scroll positions.\n- Verified on `origin/main`.", "file": "beans/defs/folio-assistant-afu3--search-pinned-the-magnifier-stays-on-screen-at-eve.md" }, { "id": "folio-assistant-ahab", "title": "Translated pages carry 225 unresolved links — 43 distinct, five near-copies each, and the English sources do not carry them", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-26T10:40:05Z", "updatedAt": "2026-09-26T17:29:28Z", "preview": "Found 2026-09-26 while re-deriving `hloc`'s premise. `hloc` records the `docs/`\nunresolved-link count as **120** after `mi97` took it from 246. Re-measured\ntoday with `scanSubgraphs()` it is **340** — higher than the pre-`mi97` figure.\n\nThat is not an `mi97` regression. It is a different population, and it splits:\n\n| source | links |\n|---|---|\n| translated pages — `docs/{ar,es,fr,ru,zh}/` | **225*", "file": "beans/defs/folio-assistant-ahab--translated-pages-carry-225-unresolved-links-43-dis.md" }, { "id": "folio-assistant-ahvw", "title": "PROCESS: how an agent decides what it is doing, and what governs each step", "status": "in-progress", "type": "epic", "priority": "normal", "parent": "folio-assistant-rwmf", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T11:43:44Z", "updatedAt": "2026-10-20T09:52:00Z", "preview": "How an agent works out what it is doing, and what governs the step it is on.\n\n`hylg` makes the workflow interpreter descend into call activities, so a\nsubprocess gives increasing context rather than a flat step name. `3190` makes\ncoding conventions a process-SCOPED context kind, so a rule applies where it\nbelongs instead of everywhere.\n\n`xfoh` and `vjbl` are the measurement half: CRDM detection ha", "file": "beans/defs/folio-assistant-ahvw--process-how-an-agent-decides-what-it-is-doing-and.md" }, { "id": "folio-assistant-ajx9", "title": "WHO IG CHROME: the blue bar, DRAFT watermark and publish box, ingested from the template chain rather than transcribed", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-yj32", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T19:56:03Z", "updatedAt": "2026-09-26T03:54:14Z", "preview": "Owner, 2026-09-23, chose \"All three, fetch HL7's base template\" and \"smart-base, inherited\".\n\nFinishes what 0818 left open: the WHO visual styling, a different question from the menu it shipped.\n\n## The template chain is three layers, and the three items do not come from one of them\n\n smart-trust local-template -> who.template.root 0.5.0 -> fhir.base.template 1.0.0\n\n- BLUE BAR is WHO's: --n", "file": "beans/defs/folio-assistant-ajx9--who-ig-chrome-the-blue-bar-draft-watermark-and-pub.md" }, { "id": "folio-assistant-akjg", "title": "DOCS-POPULATED HAS NO SUBJECT TEST: the gate passes smart-trust on docs/category/Other.md — length and authorship have teeth, 06e3 §4(b) processes/roles/tasks half was never built", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-0lmb", "blocking": [], "blockedBy": [], "createdAt": "2026-10-03T13:50:38Z", "updatedAt": "2026-10-03T14:22:22Z", "preview": "`check:docs-populated` (`scripts/check-docs-populated.ts`, 2026-10-02, gated at\n`code-quality-gates.yml:1846`) implements half of `06e3` §4(b) and reports the\nother half as if it were not asked.\n\nOwner, `06e3` §4(b), both halves in one sentence:\n\n> does the page reference the harness's **processes, roles and tasks**, and does\n> it say something about them that the index does not?\n\nBuilt: **length*", "file": "beans/defs/folio-assistant-akjg--docs-populated-has-no-subject-test-the-gate-passes.md" }, { "id": "folio-assistant-amom", "title": "MODE: test/swarm — a stack of models over shared or independent work queues", "status": "todo", "type": "feature", "priority": "normal", "parent": "folio-assistant-5a3l", "blocking": [], "blockedBy": [], "createdAt": "2026-09-19T08:55:37Z", "updatedAt": "2026-09-19T08:55:37Z", "preview": "From [#363](https://github.com/litlfred/folio-assistant/issues/363): \"test/swarm\nmode: may have a stack of models, to queue work on (a cross-model swarm mode).\nshared or indepedent work queues.\"\n\n## What exists\n\n`skills/sdlc/sdlc-core/swarm-management.md` and `docs/swarm-management.md`. The rule\nrecorded in `AGENTS.md` is that a swarm is **asked for every time, per swarm,\nwith agent count, model l", "file": "beans/defs/folio-assistant-amom--mode-testswarm-a-stack-of-models-over-shared-or-in.md" }, { "id": "folio-assistant-ansc", "title": "LSI: Latent Semantic Indexing methodology, KG index tool, need-an-LSI audit, and epic filing (#1482)", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-29T17:34:06Z", "updatedAt": "2026-09-29T23:50:43Z", "preview": "Issue #1482. Methodology node, applicability analysis, options, prototype on three WHO IRIS docs, QA audit for needing an LSI kg, then epic filing of beans/PRs/branches. Branch claude/brave-hawking-511rrx.\n\n\n## Todo\n- [x] Methodology node methodologies/lsi.md (sources located, unreachable: egress 403)\n- [x] Engine content/pipeline/lsi.ts (log-entropy + seeded randomized SVD)\n- [x] CLI scripts/lsi.", "file": "beans/defs/folio-assistant-ansc--lsi-latent-semantic-indexing-methodology-kg-index.md" }, { "id": "folio-assistant-apcg", "title": "Placement PR6: library/ingestion split — basic flow stays in the harness, l1-document-ingestion goes to core", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-iirv", "blocking": [], "blockedBy": [ "folio-assistant-63wl", "folio-assistant-b8de", "folio-assistant-tlat" ], "createdAt": "2026-10-01T06:58:01Z", "updatedAt": "2026-10-07T17:37:00Z", "preview": "Placement proposal §2 (owner rulings 2026-09-30 §6; process: review → resolve issues → staged PRs). Parent of PR0/PR1 is `9umr`; PR2–PR9 sit under the separation epic because D4 (owner, 2026-10-01, option 3) interleaves them with the split stages into one ordered sequence. Every PR: references move with the subject; generated trees regenerated, never hand-edited; `beans/**` not rewritten; kg-qa si", "file": "beans/defs/folio-assistant-apcg--placement-pr6-libraryingestion-split-basic-flow-st.md" }, { "id": "folio-assistant-apui", "title": "INGEST: one pipeline entry point — uploads/ to library/ through a single documented path", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-slw1", "blocking": [], "blockedBy": [], "createdAt": "2026-09-16T06:43:50Z", "updatedAt": "2026-10-07T17:18:00Z", "preview": "## What\n\n`uploads/` and `library/` are two stages of ONE pipeline, but there is no single\nentry point that takes a file across. Today the move is done by\n`scripts/migrate-uploads-to-library.py` plus whatever the ingesting agent\nremembers to run.\n\n## Why it matters more than it looks\n\n**The corpus-grep checklist searches `library/` only.** Anything still in\n`uploads/` is invisible to every \"has the", "file": "beans/defs/folio-assistant-apui--ingest-one-pipeline-entry-point-uploads-to-library.md" }, { "id": "folio-assistant-ar1s", "title": "Root de-pollution before the repo split: tooling configs into cat-harness-tools; root keeps only what git, agents, licences and instantiation require", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-7x5n", "blocking": [], "blockedBy": [], "createdAt": "2026-10-06T18:19:05Z", "updatedAt": "2026-10-07T17:47:00Z", "preview": "Owner ruling 2026-10-06 (session_012qoycyCSGidZqW245vXhze): 'this needs to be addressed before repo split'. Most root files are monorepo tooling, not the platform.\n\n## Target root\npackage.json + bun.lock (minimal workspace stub — bun needs both at the workspace root), .gitignore/.gitattributes/.gitmodules, AGENTS.md/CLAUDE.md/GEMINI.md, .mcp.json, .beans.yml, LICENSE/NOTICE/LICENSE-CONTENT.md/THIR", "file": "beans/defs/folio-assistant-ar1s--root-de-pollution-before-the-repo-split-tooling-co.md" }, { "id": "folio-assistant-atlf", "title": "BEANS PAGE LIGHT MODE: the selected type chip rendered black on black — it read an undefined --fa-wp-bg", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T10:45:39Z", "updatedAt": "2026-10-02T20:43:21Z", "preview": "Owner screenshot 2026-10-02 of https://litlfred.github.io/folio-assistant/beans/ with type: epic selected: the active chip is a blank black pill in light mode.\n\n## Cause\nwork-plan.css .fa-workplan-type-btn.is-active set color: var(--fa-wp-bg, #0d0d0d). Nothing defines --fa-wp-bg, so the fallback #0d0d0d painted on the light scheme's ink #0b0b0b.\n\n## Done when\n- [x] use --fa-wp-surface (defined in ", "file": "beans/defs/folio-assistant-atlf--beans-page-light-mode-the-selected-type-chip-rende.md" }, { "id": "folio-assistant-ax6r", "title": "Workflow index page is hand-maintained: generate every-workflow-in-the-repo from the process KG, strip drift; aggregate across KGs", "status": "completed", "type": "feature", "priority": "normal", "parent": "folio-assistant-whlc", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T20:42:54Z", "updatedAt": "2026-10-16T10:10:00Z", "preview": "Owner, 2026-10-02: \"why is the page hand maintained? cat-harness/content/docs/publication-workflow/every-workflow-in-the-repo.md — it should be dynamic loaded from json(ld) KG. the page itself is also a mess. lots of content drift, editorializing, PM status updates and such … once separated, this will need to be aggregated across KGs\".\n\nToday the page is a hand-written table: every new BPMN needs ", "file": "beans/defs/folio-assistant-ax6r--workflow-index-page-is-hand-maintained-generate-ev.md" }, { "id": "folio-assistant-ay3x", "title": "VECTOR FIGURE ARM: render vector drawings, and the role threshold nobody may pick from this corpus", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-2yyh", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T07:19:35Z", "updatedAt": "2026-10-03T17:46:33Z", "preview": "Issue: https://github.com/litlfred/folio-assistant/issues/877 — the deferred\nhalf of the owner's 2026-09-23 ruling on `m4xy`: **captions now, arm later.**\nThe caption half shipped in #1012.\n\n## What is deferred, and what is NOT\n\n**Deferred:** an arm that renders vector drawing groups to raster so a reader\ngets the figure itself rather than a description of it.\n\n**Not deferred, and already done:** ", "file": "beans/defs/folio-assistant-ay3x--vector-figure-arm-render-vector-drawings-and-the-r.md" }, { "id": "folio-assistant-b5f0", "title": "INSTANTIATION: what it means to instantiate a harness — the config file, the slot, and the process that cannot be started", "status": "todo", "type": "task", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T15:23:54Z", "updatedAt": "2026-10-21T19:38:00Z", "preview": "Owner, 2026-09-20, across three messages:\n\n> \"migration to cat-harness.config.json as an instance declaration in root.\n> also need bootstrap.config.json to declare its instance data upon\n> initialization. this should be part of ALL harness initializations. (not\n> dependency, but initiation = create UI/start workflow beans todos, install\n> directories etc. to a working state). this is a 'Working KG", "file": "beans/defs/folio-assistant-b5f0--instantiation-what-it-means-to-instantiate-a-harne.md" }, { "id": "folio-assistant-b7wh", "title": "INGEST arXiv:2607.14456v1 (specialist BPMN-compiled agents) as evidence + a skill on the processes subgraph", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-slw1", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T22:25:37Z", "updatedAt": "2026-09-24T00:15:16Z", "preview": "Owner, 2026-09-23: *\"ingest into library for use as evidence and skills as\nsubgraph on processes in cat-harness\"*.\n\nBorman, Wandabwa, Yu, Kannangara, Liu, Leontjeva and Ng, **\"Beyond Generalist\nLLMs: Specialist Agentic Systems for Structured Code Workflow Execution\"**,\nCommonwealth Bank of Australia. Workshop paper at SCALE, ICML 2026 (PMLR 306);\narXiv:2607.14456v1 [cs.SE], 16 July 2026. Open acce", "file": "beans/defs/folio-assistant-b7wh--ingest-arxiv260714456v1-specialist-bpmn-compiled-a.md" }, { "id": "folio-assistant-b8de", "title": "Merge refused: #1898 owed CI not green on its head", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-iirv", "blocking": [ "folio-assistant-apcg" ], "blockedBy": [], "createdAt": "2026-10-04T14:30:43Z", "updatedAt": "2026-10-07T14:55:47Z", "preview": "PR #1898, handed back by the Merge Manager on 2026-10-04 (BACKFILLED at 14:50Z: the hand-back was made as a PR comment and a queue entry only, without this bean. Owner: \"Backfill + enforce\").\n\n## Refused\n- `merge:steward` at 14:30Z: Rule_HeadNotGreen: conflict clean, own CI red.\n- Queue entry (`beans/queue/litlfred--folio-assistant--1898.json`, status `blocked`): The vqlp HOLD IS LIFTED. The owner", "file": "beans/defs/folio-assistant-b8de--merge-refused-1898-owed-ci-not-green-on-its-head.md" }, { "id": "folio-assistant-b8ip", "title": "smart-base binary artefacts: 907 MB of preview copies on gh-pages, release carries only package.tgz, downloads page promises 19 assets", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-uhkv", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T06:29:45Z", "updatedAt": "2026-10-16T10:10:00Z", "preview": "Owner, 2026-10-02: *\"see what happens on smart-base with binary artefacts on release. add to skills/tools\"*.\n\n**Measured** with `ig-binary-audit` on `litlfred/smart-base` `gh-pages` `64ed896d`:\n- **Size:** 5,029.7 MB in all, of which 966.5 MB is Publisher binaries.\n- **Previews:** 907.1 MB of those binaries (94 %) sit in 29 `branches/` previews.\n- **Largest single file:** `full-ig.zip` takes 644.7", "file": "beans/defs/folio-assistant-b8ip--smart-base-binary-artefacts-907-mb-of-preview-copi.md" }, { "id": "folio-assistant-b91x", "title": "Decide: keep beans (hmans) or move to Beads (gastownhall/beads)", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-8jt6", "blocking": [], "blockedBy": [], "createdAt": "2026-09-23T22:03:47Z", "updatedAt": "2026-09-23T22:32:13Z", "preview": "Raised 2026-09-23 while writing cat-harness/docs/prior-art.md (bean vq8o). hmans/beans has had no commit since 2026-04-06 (v0.4.2, pre-1.0); Beads is very active (v1.3.0, atomic claim, leases) but stores in Dolt and has had a retracted release and data-loss/merge issues after its SQLite->Dolt migration (gastownhall/beads #2251, #4259, #4356). No importer between the two exists. Owner decision; not", "file": "beans/defs/folio-assistant-b91x--decide-keep-beans-hmans-or-move-to-beads-gastownha.md" }, { "id": "folio-assistant-b94c", "title": "Cross-platform support: detect Linux-specific assumptions and provide a Windows path", "status": "completed", "type": "task", "priority": "high", "parent": "folio-assistant-1xhc", "blocking": [], "blockedBy": [], "createdAt": "2026-09-21T16:15:47Z", "updatedAt": "2026-10-07T17:18:00Z", "preview": "Issue [#735](https://github.com/litlfred/folio-assistant/issues/735). CRDM\nPhase 1 (needs assessment) — requester @costateixeira, sign-off @litlfred.\n\n## What\n\nA stated cross-platform policy, a mechanical check that detects Linux-specific\nassumptions across scripts, tooling, docs and content pipelines, and a Windows\npath wherever the check fires. Where Windows cannot be supported, a **declared**\ne", "file": "beans/defs/folio-assistant-b94c--cross-platform-support-detect-linux-specific-assum.md" }, { "id": "folio-assistant-b963", "title": "COLD START: the entry documents named scripts/ for the whole split, and nothing checks a command path", "status": "completed", "type": "bug", "priority": "normal", "parent": "folio-assistant-ahvw", "blocking": [], "blockedBy": [], "createdAt": "2026-09-20T18:05:19Z", "updatedAt": "2026-10-10T06:15:00Z", "preview": "## The register — the shapes have NAMES, not numbers\n\n*2026-09-20T23:1xZ, on the owner's instruction (\"2 resolve\").*\n\nThe ordinals collided: this bean numbered **shapes** in discovery order, while\n`7iog` called itself \"a fourth class beyond `b963`'s three\" — numbering\nsomething else entirely, the **readers** `check:command-paths` had at the\ntime. Two incompatible schemes, one word. And by the time", "file": "beans/defs/folio-assistant-b963--cold-start-the-entry-documents-named-scripts-for-t.md" }, { "id": "folio-assistant-ba9e", "title": "Generated README subdirectory counts: 207 volatile integers across 54 READMEs — move them to the _data layer", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-hfag", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T22:34:28Z", "updatedAt": "2026-10-04T12:13:31Z", "preview": "Child of `y7b3` part 2, which the owner left as \"decide later\" on 2026-09-30\nwith three options on the table (band the count, drop it, keep it exact).\n**Owner chose a fourth on 2026-10-02: pull the count dynamically from the KG\nJSON-LD rather than freeze it into committed markdown.**\n\n## Why that is the right answer and the other three are not\n\nThe repository had already reached this conclusion fr", "file": "beans/defs/folio-assistant-ba9e--generated-readme-subdirectory-counts-207-volatile.md" }, { "id": "folio-assistant-bamf", "title": "fhir-harness declares and populates FHIR IG Jekyll data (site.data.fhir.*) for the just-the-docs pipeline", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-zzmr", "blocking": [], "blockedBy": [], "createdAt": "2026-09-30T10:16:18Z", "updatedAt": "2026-09-30T20:09:40Z", "preview": "Owner 2026-09-30: IG pages use Liquid site.data.fhir.* (IG Publisher convention, e.g. {{site.data.fhir.ig.version}}). Those should be available in the just-the-docs pipeline, with fhir-harness RESPONSIBLE for declaring and populating the FHIR metadata Jekyll data.\n\n## Measured 2026-09-30 (issue #1564 analysis)\n- Nothing in folio-assistant resolves site.data.fhir.*: only the IG Publisher does. dak-", "file": "beans/defs/folio-assistant-bamf--fhir-harness-declares-and-populates-fhir-ig-jekyll.md" }, { "id": "folio-assistant-bbv3", "title": "Page bytes: the fa-translation-index island is fetched once, not inlined 2356 times (22.34 MB); kg-render.js carries R4's two client-side obligations", "status": "completed", "type": "task", "priority": "normal", "parent": "folio-assistant-o3xy", "blocking": [], "blockedBy": [], "createdAt": "2026-10-02T17:50:36Z", "updatedAt": "2026-10-07T20:30:00Z", "preview": "## What and why\n\n`#fa-translation-index` was a `