Glossary: Schema fields, S–Z
Candidate terms extracted from every schema field with a doc comment: <Declaration>.<field> as the label, the comment’s first paragraph as the definition. Each is the asset’s own text, verbatim and not curated, and carries the badge “candidate, extracted”. A person promotes one by authoring it. Authored terms are on their owning instance’s page, and those pages, the counts and the sources are linked from the glossary index.
From: bootstrap-tools 44 (SKOS) · cat-harness 2141 (SKOS) · fhir-harness 78 (SKOS) · folio-assistant-core 179 (SKOS) · folio-assistant-sci 13 (SKOS) · smart-base 13 (SKOS).
One of 4 pages of this type, split by the first letter of the label: all parts.
Size: this page holds 674 terms and is 366 KB before compression, fetched in one request, within its budget of 1.0 MB. There is no search index: the filter below runs over this page, and the A–Z bar jumps within it.
| target | mapped | unmapped | undetermined | why undetermined |
|---|---|---|---|---|
fhir | 0 | 2468 | 0 | — |
skos | 0 | 2468 | 0 | — |
Undetermined is never “no match”. A vocabulary that could not be reached has said nothing, and the column above keeps that apart from a checked miss. The counts are reported and not graded: an unmapped candidate may be a term this corpus is right to coin.
Per term. Every term on this page is unmapped on fhir, and unmapped on skos, unless its entry says otherwise. No entry says otherwise.
674 shown
S
- ScenarioLayer.dir candidate, extracted
Absolute path of its
scenariosdirectory.- ScenarioLayer.instance candidate, extracted
The instance's declared name.
- ScoredKeywordSchema.score candidate, extracted
Comparable within one library's matrix only.
- ScriptEntry.command candidate, extracted
The command, relative to the checkout root.
- ScriptEntry.manifest candidate, extracted
The manifest it is declared in, repository-relative (
package.jsonfor the root).- ScriptQaReport.$schema candidate, extracted
Schema marker — value:
script-qa/v1.- ScriptQaReport.criteria candidate, extracted
Per-criterion reviewer entries. Reuses the block-QA
QaCriterionEntryshape (script_hash / script_commit_sha / deps_hash on the reviewer apply identically here).- ScriptQaReport.language candidate, extracted
Source language inferred from extension.
- ScriptQaReport.references candidate, extracted
**Proper-metadata references** the script declares to propositions, definitions, theorems, glossary entries, or bibliographic keys. Mirrors the content-block convention of
uses: [...]/meta.cites: [...]arrays in.tsmanifests.- ScriptQaReport.script_path candidate, extracted
Repo-relative path to the audited script.
- ScriptQaReport.source_hash candidate, extracted
12-char SHA-256 prefix of the script's content.
- ScriptQaReport.updated_at candidate, extracted
ISO-8601 UTC timestamp of last write.
- Section.blocks candidate, extracted
Ordered list of block root names (resolved to .ts files in section/chapter dir).
- Section.lead candidate, extracted
A chapter's LEAD: the text between the chapter heading and its first section (an executive summary has nothing else). It keeps a title, for outlines and the review page, but renders without a heading of its own, because the chapter heading directly above already says it.
- Section.subsections candidate, extracted
Subsections (inline or referenced).
- SectionRef.name candidate, extracted
Root name of the section dir/file.
- SectionRef.uri candidate, extracted
URI override (defaults to
./<name>/<name>.ts).- SectionVerdictFileSchema.verdicts candidate, extracted
entry slug → section id (the
sections/<id>.mdstem) → verdict.- SectionVerdictSchema.saw candidate, extracted
What the inspector SAW — the reason for the role, checkable against the page.
- SemanticZoomSchema.belowPx candidate, extracted
The folio's default, in CSS pixels. Required — so nothing is ever unstated.
- SemanticZoomSchema.byKind candidate, extracted
Overrides, keyed by kind. A kind absent here takes the default.
- SessionContextSchema.actor candidate, extracted
WHO is acting. See the module note — this is the field the record is for.
- SessionContextSchema.claimed candidate, extracted
Beans this session has claimed.
- SessionContextSchema.id candidate, extracted
Stable for the life of the session.
- SessionContextSchema.open candidate, extracted
Instances this session has open. EMPTY IS THE NORMAL STATE, not a gap: an agent answering a question, reading, or deciding what to do next is in no process at all, and
process-statecalls that idle rather than broken.- SessionContextSchema.startedAt candidate, extracted
When the session began, as an ISO-8601 UTC instant.
- SessionContextSchema.updatedAt candidate, extracted
When this record was last written.
- SessionContextSchema.waitingOn candidate, extracted
What the session is waiting on, if anything.
- SessionHook.commands candidate, extracted
Commands to execute when the event fires.
- SessionHook.event candidate, extracted
The event that triggers this hook.
- SessionHook.matcher candidate, extracted
Tool name regex matcher (for PostToolUse).
- SessionSurveySchema.axes candidate, extracted
Per-axis findings. Empty is refused: a survey that covered nothing is not a survey.
- SessionSurveySchema.branch candidate, extracted
The branch the window is on, so a reader knows which history
from/tolive in.- SessionSurveySchema.by candidate, extracted
Who took it, so a reader can ask. A session URL or id.
- SessionSurveySchema.commits candidate, extracted
How many commits the window held, as counted at publication.
- SessionSurveySchema.from candidate, extracted
The window's LOWER edge — the last commit the surveyor already knew. Exclusive, matching
git rev-list from..to.- SessionSurveySchema.takenAt candidate, extracted
When it was taken, ISO-8601. Not the staleness test —
tois — but it dates the prose.- SessionSurveySchema.to candidate, extracted
The window's UPPER edge — the newest commit the survey actually read. Inclusive.
- ShapeField.type candidate, extracted
The type as written in the source.
- ShardSchema.entries candidate, extracted
[id, title, url], sorted by id. Arrays rather than objects: the keys would be most of the bytes.- SheetAnchorSchema.cell candidate, extracted
A1-style reference, when the format has one.
- SidecarOverlaySchema.codeCount candidate, extracted
Number of codes, for a ValueSet.
- SidecarOverlaySchema.displays candidate, extracted
schemas/<name>.displays.json— display strings, keyed by code.- SidecarOverlaySchema.jsonld candidate, extracted
<name>.jsonld— the JSON-LD vocabulary, published at the IG ROOT, not underschemas/.- SidecarOverlaySchema.openapi candidate, extracted
schemas/<name>.openapi.json— this ARTEFACT's OpenAPI fragment. Neveropenapi/openapi.json; see the module doc.- SidecarOverlaySchema.propertyCount candidate, extracted
Number of properties, for a logical model.
- SidecarOverlaySchema.schema candidate, extracted
schemas/<name>.schema.json— the JSON Schema for this artefact.- SignatureSchema.format candidate, extracted
How to interpret
value— a media type or a named scheme.- SignatureSchema.signer candidate, extracted
Who signed, as a key identifier.
- SignatureSchema.value candidate, extracted
The signature itself, or a URI that resolves to it.
- SignatureSchema.verifiedAt candidate, extracted
When it was last CHECKED against the signer's key — never when it was recorded. Same discipline as {@link FixitySchema.verifiedAt}.
- SignOffSchema.actor candidate, extracted
The declared actor this signer acted as, when the harness declares actors.
- SignOffSchema.evidence candidate, extracted
Where the decision can be read: a link to the comment it was made in.
- SignOffSchema.id candidate, extracted
Stable identifier — a login, an agent's name, a script path.
- SignOffSchema.reason candidate, extracted
Why — the dispensation, in the signer's words or quoting them.
- SignOffSchema.scope candidate, extracted
The set's id, or a member's reference.
- SignOffSchema.stage candidate, extracted
The stage this sign-off moves the set to, when its scope is the set.
- SimulatorBlock.defaultView candidate, extracted
Default parameter values — used for snapshot generation.
- SimulatorBlock.html candidate, extracted
Path to the standalone HTML simulator file (relative to repo root).
- SimulatorBlock.views candidate, extracted
Additional named views with different parameter sets.
- SimulatorRef.ref candidate, extracted
Label of the simulator block (e.g. "sim:q-double-slit").
- SimulatorRef.view candidate, extracted
Named view to use (defaults to "default").
- SimulatorView.name candidate, extracted
View name (e.g. "default", "bach-2013", "fullerene").
- SimulatorView.params candidate, extracted
Parameter values for this view.
- SimulatorView.title candidate, extracted
Display title for this view.
- SkillCapabilityRef.capabilityId candidate, extracted
The capability this skill needs.
- SkillCapabilityRef.degradation candidate, extracted
What to do when the capability is unavailable.
- SkillDefinition.allowedTools candidate, extracted
Tools the agent is allowed to use when this skill is active.
- SkillDefinition.dependsOn candidate, extracted
Dependencies on other skills or requirements.
- SkillDefinition.description candidate, extracted
What this skill does.
- SkillDefinition.id candidate, extracted
Unique skill identifier.
- SkillDefinition.lifecycleStages candidate, extracted
Lifecycle stages this skill participates in.
- SkillDefinition.name candidate, extracted
Human-readable name.
- SkillDefinition.package candidate, extracted
External package this skill belongs to (undefined = local).
- SkillDefinition.requiredCapabilities candidate, extracted
Capabilities this skill needs to function.
- SkillDefinition.routingPatterns candidate, extracted
Regex patterns for routing user requests to this skill.
Searchable tags.
- SkillDependency.conformance candidate, extracted
How strongly this dependency is required.
- SkillDependency.kind candidate, extracted
Whether the dependency is on a skill or a requirement.
- SkillDependency.ref candidate, extracted
Target skill ID or requirement ID.
- SkillPackage.mode candidate, extracted
Package mode. Defaults to "sync". - "sync": Files are copied into .claude/skills/<name>/ by the nightly workflow. - "reference": Package is registered but NOT synced. Skills are fetched on demand at runtime via the skill_fetch MCP tool.
- SkillPackage.name candidate, extracted
Local directory name under .claude/skills/.
- SkillPackage.path candidate, extracted
Path within the repository where skill directories live.
- SkillPackage.ref candidate, extracted
Git ref (branch, tag, or SHA) to sync from. Defaults to "main".
- SkillPackage.repo candidate, extracted
GitHub repository in owner/repo format.
- SkillPackage.skills candidate, extracted
List of skill directory names to sync from the package.
- SkillPackageRef.manifestPath candidate, extracted
Path to the package's
PackageManifestfor Docker deps.- SkillPackageRef.name candidate, extracted
Package name (matches directory under
.claude/skills/).- SkillPackageRef.path candidate, extracted
Path within the source repo.
- SkillPackageRef.ref candidate, extracted
Git ref to sync from (commit hash preferred).
- SkillPackageRef.repo candidate, extracted
Source repository (e.g., "K-Dense-AI/claude-scientific-skills").
- SkillPackageRef.skills candidate, extracted
Skill IDs provided by this package.
- SkillProvenance.depth candidate, extracted
Depth in the
inheritswalk; 0 is the role itself.- SkillProvenance.via candidate, extracted
The role that declared it — may be an ancestor, not the role asked for.
- SkillRegistry.actors candidate, extracted
All actor/role definitions.
- SkillRegistry.capabilities candidate, extracted
All capability definitions.
- SkillRegistry.hooks candidate, extracted
Lifecycle hooks.
- SkillRegistry.packages candidate, extracted
Skill packages under
skills/, with their Docker requirements.- SkillRegistry.repository candidate, extracted
Repository identifier (e.g., "litlfred/qou").
- SkillRegistry.requirements candidate, extracted
All workflow requirements.
- SkillRegistry.roleAssignments candidate, extracted
Identity-source → actor mapping rules, evaluated by priority (highest first).
- SkillRegistry.schemaVersion candidate, extracted
Schema version for forward compatibility.
- SkillRegistry.skills candidate, extracted
All skill definitions (local + packaged).
- SkillSchemaRef.access candidate, extracted
Whether this skill reads or writes these types.
- SkillSchemaRef.module candidate, extracted
Schema module (e.g., "schemas/types", "schemas/formalization-types").
- SkillSchemaRef.types candidate, extracted
Type name(s) within the module.
- SkillsDirectoryEntrySchema.subgraph candidate, extracted
Marks the entry as a subgraph of the INSTANCE rather than a part of the skills graph. Replaced
dependentson 2026-09-30 (option A), when inheritance became automatic and the per-entry field was retired.- SkillTopicSchema.path candidate, extracted
One plain segment: a topic is a direct subdirectory, never deeper.
- SkillTopicsSchema.directories candidate, extracted
One node, two lists: the instance directories declared from within.
- Snap.fields candidate, extracted
A contributed (structured) kind's manifest fields, read as text (bean f327). Never serialised.
- Snap.mdPath candidate, extracted
Absolute path of the prose sidecar, when there is one. Never serialised.
- SourcedEntry.storage candidate, extracted
#1764's field, if a declaration carries it. Read only to map it.
- SourceDescriptorSchema.character candidate, extracted
The characterisation the "ingest a new data source" process produces. Required — these are what decide how the source may be REFERENCED, and a descriptor that has not said cannot answer that question.
- SourceDescriptorSchema.nodeKind candidate, extracted
What one node IS in this corpus. Different per source, and never assumed.
- SourceDescriptorSchema.subsetBasis candidate, extracted
Why
subsetIsSelfContainedhas the value it has. Optional, because a self-contained source (IRIS: fetch an item and you have it) has little to say; thefalsecase is the one that needs its reason recorded.- SourceDescriptorSchema.subsetIsSelfContained candidate, extracted
Whether a subset chosen by the strategies above stands on its own.
falsemeans every request must be closed over dependencies first.- SourceDescriptorSchema.system candidate, extracted
The software, where there is some —
DSpace 7,Lean 4 / Lake. Descriptors for one system are often reusable.- SourceDescriptorSchema.termsUrl candidate, extracted
Whether the source states terms, and where. Feeds the
restrictionsandcopyrightgates;unknownstays unknown.- SourceProvenanceSchema.local candidate, extracted
The original **in this repository** this was taken from.
- SourceProvenanceSchema.upstream candidate, extracted
The remote thing. A URI, always — never a path, never a bare name. Absent means NOT RECORDED, which is not "there is none".
- SourceSchema.batch candidate, extracted
The intake batch: a file name, a form export, a letter.
- SourceSchema.entry candidate, extracted
The log's OWN number for the row ("No."), within its sheet. A review log is re-sent as it is worked, so a later copy is a different file with the same rows: this, with the sheet and the series, is what recognises them.
- SourceSchema.segment candidate, extracted
For a narrative: which passage of the letter this comment is.
- SourceSchema.series candidate, extracted
The log this row belongs to, across the copies of it that are re-sent.
- SourceSchema.sha256 candidate, extracted
sha256 of the intake file, so a re-import is recognised.
- SourceSchema.url candidate, extracted
For a GitHub comment: its URL.
- SpecValidationResult.featureName candidate, extracted
Extracted feature title/name if found.
- SpecValidationResult.findings candidate, extracted
Descriptive findings for any rule breach.
- SpecValidationResult.hasActionRequired candidate, extracted
Whether the spec carries an ACTION REQUIRED marker.
- SpecValidationResult.isAdjudicated candidate, extracted
Whether the status is considered adjudicated.
- SpecValidationResult.missingSections candidate, extracted
Any mandatory sections that were not found in the spec.
- SpecValidationResult.state candidate, extracted
"pass" if compliant, "fail" if mandatory section missing or invalid adjudication.
- SpecValidationResult.status candidate, extracted
Extracted status string from header, if found.
- SpecValidationResult.unresolvedClarifications candidate, extracted
Any unresolved [NEEDS CLARIFICATION: ...] markers.
- StageSkillBinding.skills candidate, extracted
Skills that run during this stage, with conformance.
- StageSkillBinding.stage candidate, extracted
The lifecycle stage.
- StageTransition.from candidate, extracted
Stage transitioning from.
- StageTransition.gates candidate, extracted
Gate conditions that must be met (requirement statement keys).
- StageTransition.to candidate, extracted
Stage transitioning to.
- StageTransition.trigger candidate, extracted
What triggers this transition.
- StagingComparison.branch candidate, extracted
Feature branch name (e.g. "claude/206-staging-preview").
- StagingComparison.branchUrl candidate, extracted
Full URL to the branch on GitHub.
- StagingComparison.mainBaseUrl candidate, extracted
Base URL for the main site.
- StagingComparison.missing candidate, extracted
What could not be linked, said rather than hidden: no preview yet, no page list computed, or a changed page the publish ref does not hold.
- StagingComparison.pages candidate, extracted
Individual page comparisons.
- StagingComparison.pr candidate, extracted
PR number if one exists.
- StagingComparison.slug candidate, extracted
Branch slug used in the staging URL path.
- StagingComparison.stagingBaseUrl candidate, extracted
Base URL for the staging preview.
- StagingDeployFactsSchema.branch candidate, extracted
Branch the preview was built from.
- StagingDeployFactsSchema.builtAt candidate, extracted
ISO-8601. When the preview was built.
- StagingDeployFactsSchema.commit candidate, extracted
Head commit at build time.
- StagingDeployFactsSchema.host candidate, extracted
WHICH HOST rendered it — never assumed to be gh-pages.
- StagingDeployFactsSchema.issue candidate, extracted
Issue the pull request is for, where it names one.
- StagingDeployFactsSchema.pr candidate, extracted
Pull request number, where the deploy was triggered by one.
- StagingDeployFactsSchema.slug candidate, extracted
The
STAGING/<slug>path segment. Derived, never reversed.- StagingPageComparison.after candidate, extracted
URL on the staging preview site. Absent for deleted pages.
- StagingPageComparison.before candidate, extracted
URL on the main (production) site. Absent for new pages.
- StagingPageComparison.summary candidate, extracted
What changed on this page (one-liner).
- StagingPageComparison.title candidate, extracted
Page title or path (e.g. "French landing page").
- StagingPreviewBlockSchema.observed candidate, extracted
Absent until the first sweep has looked at this preview.
- StagingPreviewBlockSchema.retiredOn candidate, extracted
ISO-8601. Set when the preview is removed; the node itself stays.
- StagingPreviewBlockSchema.retiredReason candidate, extracted
Why it was retired. Required in practice whenever
retiredOnis set.- StagingSweepFactsSchema.liveness candidate, extracted
The liveness verdict, as
previewLivenessreturns it.- StagingSweepFactsSchema.observedAt candidate, extracted
ISO-8601. When the sweep looked.
- StagingSweepFactsSchema.signals candidate, extracted
Which signals fired, for a reader who wants to know WHY.
- StandingCandidate.branch candidate, extracted
The branch the commit was resolved from — it must be the one the rule follows.
- StandingCandidate.verdict candidate, extracted
The commit's own combined status / checks:
greenonly when every one concluded successfully.- StandingConsentSchema.consentAt candidate, extracted
When it was given. Part of the rule's identity: a re-given rule re-asks every pin under the old one.
- StandingConsentSchema.consentBy candidate, extracted
Who gave the standing consent — recorded as
consent.byon every pin it covers.- StandingConsentSchema.evidence candidate, extracted
Where it was given, so it can be read. Optional:
consentByandconsentAtare the record.- StandingConsentSchema.follows candidate, extracted
The upstream branch whose newest commit the pin follows.
- StandingConsentSchema.require candidate, extracted
green: only a commit whose own combined status / checks are green. Absent: the branch head.- StickyBuildContext.createdAt candidate, extracted
ISO 8601. Supplied rather than read from the clock.
- StickyBuildContext.initiation candidate, extracted
Carried onto the node when initiation has reported. See {@link Initiation}.
- StickyBuildContext.page candidate, extracted
Defaults to {@link LANDING_STICKY_PAGE}.
- StickyContributionSchema._comment candidate, extracted
Free-form note for a reader of the declaration. Never rendered.
- StickyContributionSchema.body candidate, extracted
The sticky's text, markdown, written out. Mutually exclusive with
bodyFrom.- StickyContributionSchema.bodyAppend candidate, extracted
Markdown appended after the body, whichever way the body was obtained.
- StickyContributionSchema.bodyFrom candidate, extracted
Read the body from the declaring instance instead. See {@link STICKY_BODY_SOURCES}.
- StickyContributionSchema.id candidate, extracted
The sticky's id — and **the idempotency key**.
- StickyContributionSchema.links candidate, extracted
Links the sticky offers, written out.
- StickyContributionSchema.onboardingLinks candidate, extracted
Also offer the platform's four onboarding links.
- StickyContributionSchema.order candidate, extracted
See {@link STICKY_ORDER_LEADING}. Required: an absent order is a tie with everything.
- StickyContributionSchema.shape candidate, extracted
Which crop this sticky takes, overriding the one derived from its content.
- StickyContributionSchema.summary candidate, extracted
Shown where notes are listed. Derived from the body's first line when absent.
- StickyContributionSchema.text candidate, extracted
Where and how big this sticky's words are. See {@link StickyTextSchema}.
- StickyContributionSchema.theme candidate, extracted
The theme id, by reference. **No default, and that is load-bearing.**
- StickyLinkSchema.href candidate, extracted
Site-root-relative (
/guides/index.html) or absolute (https://…).- StickyLinkSchema.label candidate, extracted
The link text. Translatable — it is what a reader sees.
- StickyLinkSchema.note candidate, extracted
One line under the label, when the label alone is not enough. Translatable.
- StickyTextSchema.align candidate, extracted
Vertical placement within the box. Defaults to
center.- StickyTextSchema.box candidate, extracted
Override the art's cloud region.
- StickyTextSchema.scale candidate, extracted
Multiply the words' size.
1is normal;0.8fits more in.- SubgraphCoverageSchema.process candidate, extracted
The BPMN process that GOVERNS this directory — a diagram's basename, without
.bpmn, asdocument-ingestion.- SubgraphCoverageSchema.serialisations candidate, extracted
What produces this directory's SERIALISATIONS —
json,jsonldandschema.jsonat the directory's own URL.- SubgraphCoverageSchema.visualiser candidate, extracted
The page that renders this subgraph, **relative to the REPOSITORY root** — not to the instance, which is what the sibling
pathon {@link ContentDirectorySchema} is relative to.- SubgraphNodeBase.holdsGraph candidate, extracted
The graph typologies it holds, as bootstrap's graph-typology individuals.
- SubgraphNodeBase.holdsGraph candidate, extracted
Links to the GraphTypology nodes this subgraph holds.
- SubgraphNodeBase.name candidate, extracted
The repository's or harness's name at the two top levels; the instance-relative path below them.
- SubgraphNodeBase.name candidate, extracted
The subgraph's label: the harness name at the root, the instance-relative path below it.
- SubgraphNodeBase.path candidate, extracted
Instance-relative, ending in
/;./at the two top levels.- SubgraphRelation.children candidate, extracted
Directories nested immediately inside it, by
id.- SubgraphRelation.parent candidate, extracted
The containing directory's
id.- Subject.scheme candidate, extracted
The candidate's scheme id, as the mapping record keys it (
kg-tools, …).- Subject.term candidate, extracted
The candidate's term id within that scheme.
- Subscription.assets candidate, extracted
How referenced binary assets are materialised. Each copy still passes
Process_MaterializeRemote's gates.- Subscription.harnesses candidate, extracted
Harness names from the substrate CHOSEN for instantiation. Instantiating writes
<name>.config.json, which is what puts it in the navbar.- Subscription.id candidate, extracted
Local name for the subscription; unique within this list. The substrate's own name, unless two pins of one substrate are ever needed.
- Subscription.kind candidate, extracted
contentfor a Knowledge Graph with Subgraphs and no harness (owner, 2026-10-06); absent:substrate. See {@link SubscriptionSchema}.- Subscription.note candidate, extracted
One sentence a reader of the visualizer sees.
- Subscription.ref candidate, extracted
The pinned commit, a full 40-character SHA. A tag replaces it once the substrate publishes releases.
- Subscription.repository candidate, extracted
Where the substrate lives,
owner/repo(#1652).- Subscription.subgraphs candidate, extracted
Subgraph ids (the substrate's
directories[].id) CHOSEN for materialisation. Everything else stays referenced.- Subscription.upstreamPath candidate, extracted
The repository-relative directory holding the substrate's declaration; absent: the root. See {@link SubscriptionSchema}.
- SubscriptionSchema.kind candidate, extracted
What kind of Knowledge Graph this subscribes to —
substrate(absent) orcontent(owner, 2026-10-06): a declaration with Subgraphs and NO harness, such as a FHIR IG. Bootstrap's definition of a substrate is unchanged; a content subscription is a separate kind, written explicitly, and contributes no skills, processes or roles — so it may choose noharnesses.SUBSCRIPTION_KINDSinschemas/substrate-snapshot.tscarries the rule once.- SubscriptionSchema.upstreamPath candidate, extracted
The repository-relative DIRECTORY holding the substrate's declaration, when that is not the root —
sharedfor a fork whose declaration isshared/<name>.json(bean437w). Absent: the root. The same name and meaning asupstreamPathon a remote subgraph source (schemas/subgraph-source.ts, PR #2326), so one word means one thing across both, and the same pattern (a trailing slash is admitted and means nothing). Recorded sokg:subscribe:checkand a re-subscribe judge the same subtree that was judged the first time.- SubstrateSnapshotSchema.file candidate, extracted
The upstream file, relative to the substrate's repository root:
<name>.jsonat the root, or<upstreamPath>/<name>.jsonwhen the declaration is nested (bean437w). Its directory is the subscription'supstreamPath, andkg:subscribe:checkholds the two equal.- SubstrateSnapshotSchema.kind candidate, extracted
See {@link SUBSCRIPTION_KINDS}. Absent:
substrate. Written only ascontent.- SubstrateSnapshotSchema.raw candidate, extracted
The upstream bytes, exactly as fetched.
- SubstrateSnapshotSchema.ref candidate, extracted
The pinned commit the bytes were read at — a full SHA, never a branch.
- SubstrateSnapshotSchema.subscription candidate, extracted
The
subscriptions[].idthis snapshot belongs to.- SubstrateSnapshotSchema.summary candidate, extracted
What the judgement found, derived from
raw.- SuccessCriterionSchema.criterion candidate, extracted
One sentence a reviewer can say yes or no to, without asking what it means.
- SuccessCriterionSchema.key candidate, extracted
Unique within its statement;
req:<id>#<statement>/<key>is its address.- SurveyAxisSchema.axis candidate, extracted
What was examined —
beans,ci,prs,issues,branches, …- SurveyAxisSchema.covered candidate, extracted
False means NOT LOOKED AT;
findingthen says why, not what was found.- SurveyAxisSchema.finding candidate, extracted
What it found, or — when
coveredis false — why it was skipped.- SweepMarkSchema.by candidate, extracted
Who ran it, so a reader can ask: a session URL or id. Absent when the runner did not know.
- SweepMarkSchema.check candidate, extracted
The check this marks, by its
catcommand name —session-staleness. Also the file's stem.- SweepMarkSchema.checkedAt candidate, extracted
When the check last read its input, ISO-8601 with a zone.
- SweepMarkSchema.findings candidate, extracted
How many findings it reported.
- SweepMarkSchema.read candidate, extracted
How many subjects the run read — sessions, for
session-staleness. Zero is a real answer.
T
- TableBlock.caption candidate, extracted
Caption text (rendered above table in LaTeX).
- TableBlock.label candidate, extracted
Table label (e.g. "tbl:mass-predictions").
- TableBlock.tex candidate, extracted
LaTeX tabular/longtable source (optional; can also live in .md).
- TabularCsvwSchema.narrative candidate, extracted
What the data is about, under bean
ju0u's state machine, the same record v1 carries. Required on the same terms assource.- TabularCsvwSchema.source candidate, extracted
The technical-metadata block every ingest sidecar carries (
file,sha256,mimetype_sniffed, ...), the same recordfolio-tabular-records/v1holds assource— beanurat. It is howcheck-l1-complete(sourceBlockOf) andcheck-library-qalearn what the entry was made from. Required (urat step 5): no committed record predates it, and an extractor's output, which has none, is {@link TabularCsvwTablesSchema}. Not a CSVW key: {@link toCsvw} never emits it.- TabularCsvwSchema.tables candidate, extracted
One per sheet. A CSV has exactly one; a workbook is a
csvw:TableGroup.- TabularRecordsSchema.@context candidate, extracted
The published content context — bean
yh6u. OPTIONAL because folio repositories hold records written before the arm emitted it; when present it must be that context (bare, or with@basebeside it — beanbh4q), since any other would bind these keys to terms nobody declared.- TabularRecordsSchema.header_vocabulary candidate, extracted
Every distinct header across every sheet — what a grep lands in.
- TabularSheetSchema.headers candidate, extracted
Row 1, in column order. Empty when the sheet has no readable header row.
- TacticMapping.context candidate, extracted
Context in which this mapping applies.
- TacticMapping.fallback_tactics candidate, extracted
Fallback tactics if primary fails.
- TacticMapping.is_regex candidate, extracted
Whether the pattern is a regex.
- TacticMapping.pattern candidate, extracted
Narrative phrase pattern (regex or literal).
- TacticMapping.primary_tactic candidate, extracted
Primary Lean tactic to try.
derived: a rendering of another target's output, copied from it.- TargetSchema.code candidate, extracted
The target term: a predicate's local name, or a code in the target system.
- TargetSchema.derivedFrom candidate, extracted
The
key(orcode) of the authoritative target a derived one copies.- TargetSchema.equivalence candidate, extracted
The R4 statement, verbatim, when the row came from an R4 map.
- TargetSchema.key candidate, extracted
The JSON term written, when it is not
code.- TaskRefSchema.process candidate, extracted
BPMN process id, e.g.
Process_CRDM.- TaskRefSchema.task candidate, extracted
Activity id within that process, e.g.
A_Implement.- TermGloss.gloss candidate, extracted
One sentence. What the term IS, not why it was designed that way.
- TermGloss.layer candidate, extracted
Which layer owns it. Absent means
harness— the middle, not a guess.- TermGloss.replacedBy candidate, extracted
The standard property that replaced this term, as a CURIE over {@link REPLACEMENT_PREFIXES} — a CURIE rather than an IRI because
namespaces.tsimports this module, so the namespace constants cannot be imported back. Set when a term minted here turned out to restate a published one (owner, 2026-09-30, beanxsqm: "emphasize preexisting standards … now align"). The term stays in the vocabulary,owl:deprecatedanddcterms:isReplacedBythis IRI, so data already carrying the old IRI still resolves to a definition that says where to go — a retired IRI that 404s is a broken link someone else owns.- TermGloss.seeAlso candidate, extracted
Site-relative page carrying the full treatment, when one exists.
- TermMappingSchema.matches candidate, extracted
Present when either state is
mapped; the concepts that matched.- TermMappingSchema.scheme candidate, extracted
The scheme the candidate came from —
kg-skills,kg-tools, …- TermMappingSchema.term candidate, extracted
The candidate's id, as its glossary scheme records it.
- TermMappingSchema.undetermined_reason candidate, extracted
Required when either state is
undetermined.- TermMappingsFileSchema.scope candidate, extracted
One entry per target, always both, so a reader sees what was NOT asked.
- TermSchema.broader candidate, extracted
Local term ids or absolute IRIs.
- TermSchema.iri candidate, extracted
The concept's IRI when its defining vocabulary already mints one. The glossary then DESCRIBES that concept — its labels, translations, order — rather than minting a second IRI for the same term. Owner, 2026-09-30 (bean
xsqm, "one SKOS", option A): bootstrap's terms are<bootstrap>/ns#Node, not also…ns#glossary/terms/node.- TermSchema.isDefinedBy candidate, extracted
The schema or standard that defines the term (rdfs:isDefinedBy).
- TermSchema.notation candidate, extracted
The code. SKOS's own word for it.
- TermSchema.reason candidate, extracted
Required for
could-not-extract: a person must be able to act on it.- TermSchema.requires candidate, extracted
Local ids of the terms this definition USES — what a reader must know first (dcterms:requires). In an
orderedglossary each must come earlier; the scheme refuses one that does not.- TermSchema.source candidate, extracted
Where the term came from: a repository path (with #anchor) or an IRI.
- TestAssertionSchema.expectation candidate, extracted
What is expected, for a reader who has only the plan.
- TestAssertionSchema.type candidate, extracted
FHIR
TestPlan.testCase.assertion.type: does a failure fail the case?- TestCaseVerdictSchema.assertions candidate, extracted
Per-assertion results, keyed by the assertion's criterion id.
- TestCaseVerdictSchema.reason candidate, extracted
Why the case was skipped. Required when
resultisskipped.- TestCaseVerdictSchema.reviewed_at candidate, extracted
ISO-8601 UTC.
- TestDefinition.description candidate, extracted
Human-readable description.
- TestDefinition.file candidate, extracted
File path relevant to this test (relative to repo root).
- TestDefinition.id candidate, extracted
Unique test ID (e.g., "lean-compile:QOU.Torsion").
- TestDefinition.latex_label candidate, extracted
LaTeX label (for coverage tests).
- TestDefinition.lean_decl candidate, extracted
Lean declaration name (for compile/sorry tests).
- TestDefinition.lean_project candidate, extracted
Lean project this test applies to (e.g., "QOU", "Fred2005"). Null = repo-wide.
Tags for filtering (e.g., ["braids-and-knots", "torsion", "blocking"]).
- TestPlanCaseSchema.assertions candidate, extracted
A case with no assertion asserts nothing, and would pass vacuously.
- TestPlanCaseSchema.dependsOn candidate, extracted
Sibling cases that must run first. FHIR
testCase.dependency, resolved.- TestPlanCaseSchema.feature candidate, extracted
Optional Gherkin
.featurefile the case is written in.- TestPlanCaseSchema.testData candidate, extracted
May be empty: a case can test behaviour on no input. Absent and empty are the same here.
- TestPlanExitCriteriaSchema.decision candidate, extracted
The certification rule, executable.
- TestPlanExitCriteriaSchema.description candidate, extracted
A reader's summary of it. Never the rule itself — the DMN is.
- TestPlanRefSchema.id candidate, extracted
The plan's
id(test-plan/v1).- TestPlanRefSchema.version candidate, extracted
The plan's
versionat execution time.- TestPlanSchema.requirements candidate, extracted
What the plan is evidence FOR. At least one: a plan tied to no requirement certifies nothing.
- TestPlanSchema.version candidate, extracted
The PLAN's version — a run records which one it executed.
- TestPlanScopeSchema.target candidate, extracted
The specific system, when the plan is for one: a skill name, a tool name, an IG canonical. Absent means the plan applies to any system of
kind(a conformance plan any implementer may run).- TestPlanScopeSchema.versionRange candidate, extracted
The versions the plan applies to, as a semver range (
>=1.2 <2,*). Required:*is a statement; absence would be an omission that reads the same.- TestReport.commit_sha candidate, extracted
Git commit SHA.
- TestReport.generated_at candidate, extracted
ISO-8601 timestamp of the run.
- TestReport.lean_projects candidate, extracted
Lean project(s) tested (e.g., ["QOU", "Fred2005"]).
- TestReportSchema.cases candidate, extracted
Case id → reviewer entries, oldest first. The last is current.
- TestReportSchema.plan candidate, extracted
ONE plan. There is no field for a second — see the module note.
- TestResult.details candidate, extracted
Structured details (compiler output, diff, etc.).
- TestResult.duration_ms candidate, extracted
Duration in milliseconds.
- TestResult.message candidate, extracted
Human-readable message (especially on failure).
- TestResult.test_id candidate, extracted
References TestDefinition.id.
- TestResult.timestamp candidate, extracted
ISO-8601 timestamp.
- TestReviewerSchema.actor candidate, extracted
On whose authority — an id under
scenarios/actors/. Absent is unresolved, not permitted.- TestReviewerSchema.id candidate, extracted
What ran: a script path, an agent name, a GitHub login.
- TestRunSchema.cases candidate, extracted
The cases, when the runner records them. Absent is a real state: a run that only records aggregates cannot be checked against the contract, and
test-run-conformssays so rather than passing it.- TestRunSchema.data candidate, extracted
WHAT was tested.
- TestRunSchema.outcome candidate, extracted
The measurements themselves, in whatever shape the runner records.
- TestRunSchema.plan candidate, extracted
The plan this run executed (bean
ygzh). Optional: a run recorded before plans existed —crdm-detect-eval— tested a skill directly, and still parses.- TestRunSchema.process candidate, extracted
WHAT TESTED IT — runner, configuration, scenario bank.
- TestRunSchema.requirements candidate, extracted
WHICH REQUIREMENTS THIS RUN CHECKS — owner, 2026-09-23 (issue #1164): *"put requirements in Test schema as array"*, as REFERENCES (
req:<slug>orreq:<slug>#<statement-key>), never copies: the requirement's text lives in one place, and a copy in every run is a copy free to disagree with it.- TestRunSchema.skill candidate, extracted
The skill this run tests, by name.
- TestRunSchema.subject candidate, extracted
What this run measured, for a reader who has only the file.
- TestRunSchema.sut candidate, extracted
The system under test (bean
ygzh). Required wheneverplanis set.- TestRunSchema.updated_at candidate, extracted
ISO-8601 UTC. Deliberately NOT part of either hash: when a run happened is not what makes it reproducible, and including it would make every run differ from every other by construction.
- TestSummary.by_category candidate, extracted
Breakdown by category.
- TestSummary.duration_ms candidate, extracted
Duration of entire run in milliseconds.
- TextFileSchema.format candidate, extracted
How the rung read it, decided by CONTENT:
markdownwhen it has an ATX heading outside fenced code and does not open with<;verbatimfor any other text;imagefor a member the caller declared as an image, which is recorded and never sectioned.- TextFileSchema.lines candidate, extracted
Line count, or null for a file not read as text (an image the source carries).
- TextSectionSchema.file candidate, extracted
The file the section is in, as recorded in
source.files[].path.- TextSectionSchema.line_end candidate, extracted
One-based last line, inclusive.
- TextSectionSchema.line_start candidate, extracted
One-based first line, inclusive.
- TextSectionSchema.unit candidate, extracted
heading: divided by the file's own heading.file: the whole file, which carries none the rung reads.- TextSourceSchema.bytes candidate, extracted
Total bytes over every file.
- TextSourceSchema.file candidate, extracted
A display name for the whole source: the file's name, or the bundle's.
- TextSourceSchema.mimetype_sniffed candidate, extracted
Text has no magic bytes; what it is was decided by reading it.
- TextSourceSchema.mtime candidate, extracted
The source's own time. For a repository checkout a file's mtime is the checkout's, which says nothing, so this is the commit time instead.
- TextSourceSchema.sha256 candidate, extracted
With one file, that file's sha256. With several, the sha256 of their
sha256sum-format listing (<hex> <path>\n, infilesorder), which anyone can recompute from a checkout withsha256sumalone.- TextStructureSchema.structure_note candidate, extracted
What the rung did NOT claim. Always required: a text source always omits something (rendering, shortcodes, images).
- TextStructureSchema.toc_source candidate, extracted
headings: at least one file was divided by its own headings.files: none was, so each file is one section.- TextUpstreamSchema.committed candidate, extracted
Commit time, ISO 8601, as git reports it.
- TextUpstreamSchema.path candidate, extracted
Path of the files within the repository.
- TextUpstreamSchema.ref candidate, extracted
The ref the commit was taken from (a branch or tag name).
- TextUpstreamSchema.retrieved candidate, extracted
When it was retrieved, ISO date.
- ThemeArtFailure.detail candidate, extracted
What was found.
- ThemeArtFailure.layout candidate, extracted
The layout at fault, where the failure is about one.
- ThemeArtFailure.remedy candidate, extracted
What to do about it.
- ThemeArtIntakeInput.declaredDirectories candidate, extracted
Declared directory paths, instance-relative. The instance's own declaration supplies these.
- ThemeArtIntakeInput.destinationDir candidate, extracted
Where the files will live, relative to the instance root.
- ThemeArtIntakeInput.fileStem candidate, extracted
Extension for the emitted
src; defaults to the candidate's measured format.- ThemeArtIntakeInput.role candidate, extracted
The image role the set will be declared under, e.g.
landing-engineer.- ThemeBackdropSchema.description candidate, extracted
What the art is FOR, for an author reading the theme.
- ThemeBackdropSchema.imageRole candidate, extracted
The declared image
rolethis theme's art comes from —landingfor the threelanding-*entries an instance declares.- ThemeBackdropSchema.scrim candidate, extracted
The layer between art and ink, as a CSS colour — normally a translucent form of
surface. Required; see module docs.- ThemedTodoFieldsSchema.layout candidate, extracted
Which of the theme's three crops this sticky shows. ABSENT means the square
cardcrop, on every screen — owner, 2026-09-24: *"i want sticky themes to by default use the square avatar layout but mostly faded, if not specified."* Absent rather than defaulted, for the same reason astheme: "the author chose card" and "nobody chose" stay distinguishable.- ThemeGeometrySchema.fontScale candidate, extracted
Body scale relative to the page, as a unitless multiplier.
- ThemeGeometrySchema.minWidth candidate, extracted
Minimum column width before the grid reflows, as a CSS length.
- ThemeGeometrySchema.padding candidate, extracted
Card padding, as a CSS length.
- ThemeOverrideNoteSchema.field candidate, extracted
Dotted path into the theme, e.g.
palette.accent.- ThemeOverrideNoteSchema.reason candidate, extracted
Why it differs from the inherited value.
- ThemeOverrideNoteSchema.source candidate, extracted
Where the value came from — a CSS custom property, a voice rule, a page.
- ThemeRefSchema.instance candidate, extracted
The owning instance, as
owner/repo(bean6rmv). Absent means the citing instance's own.- ThemeSchema.backdrop candidate, extracted
Artwork behind the ink, when the theme has any.
- ThemeSchema.description candidate, extracted
One line for the theme picker. Translatable, same reasoning as
name.- ThemeSchema.inherits candidate, extracted
The theme this one starts from. Absent means it starts from nothing.
- ThemeSchema.layouts candidate, extracted
Geometry, by kind. All three after resolution, or invalid.
- ThemeSchema.name candidate, extracted
Display name.
- ThemeSchema.overrides candidate, extracted
Why each overridden value differs from the inherited one.
- ThemeSchema.palette candidate, extracted
Shared across every kind, deliberately.
- ThresholdBasisSchema.coverage candidate, extracted
Placed area over page area — the number compared against
value.- ThresholdBasisSchema.suppliedBy candidate, extracted
Always the caller. Recorded so the provenance is read, not inferred.
- ThresholdBasisSchema.value candidate, extracted
The cutoff, exactly as supplied. Images below it were filed
furniture.- TileCount.count candidate, extracted
The headline number.
0is a legitimate, meaningful value.- TileCount.unit candidate, extracted
What is being counted, as a noun the reader sees —
"entries","beans","waiting". Pluralisation is the declarer's, because only it knows whether its unit pluralises regularly.True when THIS branch, not
main, is where readers must look.- TipManifestSchema.source candidate, extracted
The commit the content was seeded from.
- TipManifestSchema.subgraph candidate, extracted
The subgraph it holds. Absent only on the retired multi-graph
statebranch.- TodoGraphSchema.defaultTheme candidate, extracted
The theme a todo takes when it does not choose one — an id from
THEMES.- TodoGraphSchema.name candidate, extracted
Display name — whose todos these are.
- TodoIndexItemSchema.target candidate, extracted
The same attachment, in parts. ABSENT when the note is attached to nothing, or when its label resolves to no block in this build — the third state, reported rather than guessed.
- TodoIndexItemSchema.targetLabel candidate, extracted
The composite label. KEPT — see the module docs.
- TodoIndexItemSchema.viewHref candidate, extracted
Present together or not at all — both come from one
sourceLinkscall.- TodoIndexSchema.repoWeb candidate, extracted
The forge address, resolved rather than composed —
gen-docs-pages'REPO_WEB.- TodoIndexSchema.tile candidate, extracted
The tile's declared headline number, keyed by directory — issue #856. Optional, because declaring one is the projection's choice and a required field would make "no count" impossible to express.
- TodoItem.assignee candidate, extracted
Who this is assigned to (agent name or "author").
- TodoItem.comment candidate, extracted
Markdown narrative with context, rationale, questions.
- TodoItem.createdAt candidate, extracted
ISO 8601 timestamp of creation.
- TodoItem.data candidate, extracted
Structured data — schema is at agent/user discretion.
- TodoItem.id candidate, extracted
Unique ID (e.g. "todo-001" or a nanoid).
- TodoItem.origin candidate, extracted
Who/what created this todo.
- TodoItem.priority candidate, extracted
Triage priority.
Labels of related todos (for threading).
- TodoItem.status candidate, extracted
Current lifecycle status.
- TodoItem.summary candidate, extracted
Short summary (one line).
- TodoItem.targetLabel candidate, extracted
Label of the block this todo is attached to (e.g. "def:rigid-monoidal-category").
- TodoItem.theme candidate, extracted
Which theme's art backs this todo's sticky — declared, never inferred.
- TodoItem.updatedAt candidate, extracted
ISO 8601 timestamp of last update.
- TodoItem.updatedBy candidate, extracted
Agent/user who last updated.
- TodoItemSchema.theme candidate, extracted
Which theme's art backs this todo's sticky.
- TodoTargetSchema.label candidate, extracted
The block's declared, page-qualified label — the same value as
targetLabel.- TodoTargetSchema.node candidate, extracted
The block's id within that page.
- TodoTargetSchema.page candidate, extracted
The page slug the block is on.
- ToolContribution.name candidate, extracted
Stable identifier, e.g.
"lean". Unique across contributors.- ToolDefinitionSchema.downstream candidate, extracted
A downstream output this Tool keeps current, with its declared inputs. See {@link ToolDownstreamSchema}. Optional because most Tools keep no derived output; a run record or a publish verifier naming a Tool that does not declare one is a finding (
downstream-tool-declared).- ToolDefinitionSchema.maintains candidate, extracted
Published artefacts this Tool is authoritative for. See {@link ToolMaintainsSchema}.
- ToolDefinitionSchema.remedies candidate, extracted
One entry per host this Tool reaches, saying what to do when it is refused. See {@link ToolRemedySchema}. REQUIRED non-empty when
requires.networkis true;check:toolsenforces it, and resolves everytoolit names.- ToolDefinitionSchema.renders candidate, extracted
The graph TYPOLOGIES this Tool draws a viewer for — one page per declared directory of the kind, placed by the Tool itself.
- ToolDefinitionSchema.satisfies candidate, extracted
Skills this Tool can satisfy. One skill may have several Tools.
- ToolDefinitionSchema.selection candidate, extracted
Why to reach for this one. REQUIRED when the Tool has a derived alternative ({@link deriveAlternatives});
check-toolsenforces it, since whether a Tool has one is a fact about the whole set.- ToolDefinitionSchema.subprocesses candidate, extracted
Process ids of the BPMN this Tool's OWN specific procedure is drawn as.
- ToolDownstreamSchema.inputs candidate, extracted
What it is derived from, so a reader can see what makes it stale.
- ToolDownstreamSchema.output candidate, extracted
What the Tool keeps current — repo-relative for
checkout, site-relative forpublished.- ToolDownstreamSchema.verifier candidate, extracted
publishedonly: thepublish-verifyverifier id that judges the output.- ToolEntrySchema.home candidate, extracted
The tool's root inside an unpacked archive, relative and POSIX.
- ToolInputSchema.arg candidate, extracted
Where this input goes when the tool is invoked. Absent means the input is part of the contract but not passed on the command line — a Tool invoked
manually orinProcesshas no command line at all.- ToolInputSchema.repeated candidate, extracted
The input is a **list** of the named type rather than one value.
- ToolInputSchema.required candidate, extracted
Whether a caller must supply it. Explicit, never inferred from a default.
- ToolInstallSchema.none candidate, extracted
Nothing to install. Stated, so "absent" and "not needed" are distinct.
- ToolInvocationSchema.argv candidate, extracted
The exact argv handed to the process. Empty for a
manualTool.- ToolInvocationSchema.inputs candidate, extracted
Parsed, type-checked input values — never the raw request.
- ToolInvocationSchema.reason candidate, extracted
Why, when the outcome was a refusal. Required for every non-
ranoutcome.- ToolInvocationSchema.skill candidate, extracted
The skill this run was exercising — one the Tool
satisfies.- ToolInvocationSchema.tool candidate, extracted
ToolDefinition.id.- ToolInvokeSchema.inProcess candidate, extracted
A function in **this instance's own code**.
- ToolInvokeSchema.manual candidate, extracted
The tool is performed by a person following the skill, with no command.
beans-manualis the case: editing a bean's front matter by hand is a real mechanism with equal standing to the CLI, and modelling it as an absentshellwould have made it indistinguishable from an unfinished record.- ToolInvokeSchema.shell candidate, extracted
The command a caller types, from the repository root.
- ToolMaintainsSchema.artefact candidate, extracted
Published path relative to the instance's base, e.g.
tool.schema.json. No leading slash: it is joined byrenderingPath, which owns the layout.- ToolMaintainsSchema.format candidate, extracted
What the artefact IS, for a consumer deciding whether it can read it. Open string: an instance may maintain kinds this vocabulary has not met.
- ToolMaintainsSchema.source candidate, extracted
Repo-relative module that defines it, e.g.
schemas/tool.ts.- ToolOutputSchema.render candidate, extracted
How a renderer may place this output in a page. Absent is
text. See {@link RENDER_AS}.- ToolProfileSchema.description candidate, extracted
What the profile is for, for a reader choosing one.
- ToolProfileSchema.tools candidate, extracted
Tool name → the release of it this profile selects. At least one: an empty profile provisions nothing and says so by not existing.
- ToolReleaseSchema.digest candidate, extracted
Required: the resolver's one refusal is a mismatch against this.
- ToolReleaseSchema.licence candidate, extracted
An SPDX licence expression, one line. Checked against the pinned list by {@link toolReleaseLicenceProblem}.
- ToolReleaseSchema.licenceSource candidate, extracted
Where the licence was read from — the jar's own
-license, the archive'slegal/, the repository'sLICENSE.- ToolReleaseSchema.note candidate, extracted
Anything a resolver or a reader must know that the fields cannot say.
- ToolReleaseSchema.runtimes candidate, extracted
Every runtime this release needs. Required, so "needs none" is stated rather than omitted.
- ToolRemedySchema.error candidate, extracted
Text an agent sees when the host refuses, if it does not name the host — matched as a substring.
- ToolRemedySchema.host candidate, extracted
The host this Tool reaches, as it appears in a refusal —
packages.fhir.org.- ToolRemedySchema.none candidate, extracted
Stated: no Tool works around this refusal, and why — or what to do instead.
- ToolRemedySchema.tool candidate, extracted
The declared Tool that does the job without this host.
check:toolsresolves it.- ToolRunRecordSchema.detail candidate, extracted
For a failure: what went wrong, for a reader who has only the file.
- ToolRunRecordSchema.inputFingerprint candidate, extracted
The tool's own fingerprint of the inputs the run read, or
unknownwhen the run failed before it could compute one.unknownmatches nothing.- ToolRunRecordSchema.target candidate, extracted
Which of the tool's outputs — one tool may keep several (LSI: one per graph).
- ToolRunRecordSchema.tool candidate, extracted
The Tool node id whose run this records. @ref ToolDefinitionSchema
- ToolSelectionSchema.cost candidate, extracted
What having it available costs. Distinct from
install; see above.- ToolSelectionSchema.limits candidate, extracted
Where it stops, and which sibling picks up — "none" is a real value, stated.
- ToolSelectionSchema.when candidate, extracted
The case this Tool is the right answer to.
- Topology.outwardFacing candidate, extracted
Whether the deployment serves people outside the operator. §1 axis 10.
- TopologyConflict.pair candidate, extracted
The two declared values that cannot co-occur, as
axis: value.- TopologyConflict.reason candidate, extracted
Why the mechanism forbids it — an entailment, never a report.
- TransitionInput.decision candidate, extracted
atandbydefault to the transition's.- TransitionInput.recommendation candidate, extracted
atdefaults to the transition's.- TranslatableFormatSchema.extensions candidate, extracted
File extensions this format applies to.
- TranslatableFormatSchema.extractModule candidate, extracted
TypeScript module that implements extraction (relative to repo root).
- TranslatableFormatSchema.id candidate, extracted
Format identifier.
- TranslatableFormatSchema.injectModule candidate, extracted
TypeScript module that implements injection (relative to repo root).
- TranslatableFormatSchema.name candidate, extracted
Human-readable name.
- TranslatableFormatSchema.notes candidate, extracted
Notes about translating THIS format specifically, as distinct from
ContentTypeTranslation.notes, which describes the content type as a whole. "Lean 4 terms stay in English" and "the diagram is re-rendered after injection" are properties of the format, not of the folio.- TranslatableFormatSchema.smartBaseExtractRef candidate, extracted
Smart-base function/line range for extraction.
- TranslatableFormatSchema.smartBaseInjectRef candidate, extracted
Smart-base function/line range for injection.
- TranslatableFormatSchema.smartBaseScript candidate, extracted
Smart-base Python script that handles this format (reference).
- TranslationBlockQaReport.block candidate, extracted
The content block this translates, by ITS label — a different identity.
- TranslationBlockQaReport.label candidate, extracted
trans:<locale>/<stem>, the conventionschemas/translation.tsfixes.- TranslationBlockQaReport.paths candidate, extracted
Repo-relative paths of the block's files.
- TranslationBlockQaReport.po candidate, extracted
The PO this verdict was measured against, repo-relative.
- TranslationConfig.defaultLocale candidate, extracted
Source language of the folio's content (BCP 47). Default: "en".
- TranslationConfig.defaultLocale candidate, extracted
BCP 47 tag for the source language. Default: "en".
- TranslationConfig.officialOnly candidate, extracted
When true, only signed-off translations are rendered. Default: false.
- TranslationConfig.officialRequired candidate, extracted
If true, only official translations are shown in the viewer. Default: false (unofficial translations are shown with a badge).
- TranslationConfig.potGranularity candidate, extracted
Whether to generate one POT per block, per chapter, or one for the whole folio.
- TranslationConfig.supportedLocales candidate, extracted
Target languages. Default: six UN languages.
- TranslationConfig.supportedLocales candidate, extracted
Locales the folio supports. Default: the six UN languages.
- TranslationConfig.translationDir candidate, extracted
Directory for .pot/.po files. Default: "translations".
- TranslationConfig.translationDir candidate, extracted
Subdirectory name within content nodes. Default: "translations".
- TranslationNode.coverage candidate, extracted
Translation coverage metrics.
- TranslationNode.description candidate, extracted
Optional description of what this PO file covers.
- TranslationNode.label candidate, extracted
Label following the
trans:<locale>/<stem>convention.- TranslationNode.locale candidate, extracted
BCP 47 locale tag of the target language.
- TranslationNode.poFile candidate, extracted
Path to the
.potranslated file (relative to folio root). This is the file that carries the actual translations.- TranslationNode.poIncludes candidate, extracted
One or more additional PO files this translation depends on. Useful for shared glossaries or terminology POs that provide base translations overridden by the primary
poFile.- TranslationNode.potFile candidate, extracted
Path to the
.pottemplate file (relative to folio root). Optional — omit for hand-authored glossary PO files that have no extractable source.- TranslationNode.sourceFile candidate, extracted
Path (relative to folio root) of the **source** file this PO translates. For content blocks, this is the
.mdsibling.- TranslationNode.sourceHash candidate, extracted
SHA-256 hash (12-char prefix) of the source file at extraction time. Used for staleness detection — if the current source hash differs, the POT is stale and needs re-extraction.
- TranslationNode.status candidate, extracted
Translation status (official/unofficial, staleness, etc.).
- TranslationNode.title candidate, extracted
Optional title for display in the viewer / docs.
- TranslationStatus.flaggedForReview candidate, extracted
Whether the translation has been flagged for review. Typically set by round-trip QA when semantic drift is detected.
- TranslationStatus.flagReason candidate, extracted
Reason for flagging, if any.
- TranslationStatus.generatedAt candidate, extracted
ISO 8601 timestamp of generation.
- TranslationStatus.generatedBy candidate, extracted
What generated the translation.
- TranslationStatus.generator candidate, extracted
The tool or model used (e.g. "DeepL", "GPT-4o", "google-translate").
- TranslationStatus.level candidate, extracted
The hierarchy level at which the sign-off was recorded.
- TranslationStatus.locale candidate, extracted
BCP 47 locale tag (must match the enclosing directory name).
- TranslationStatus.official candidate, extracted
Whether a human has signed off on this translation.
- TranslationStatus.signedOffAt candidate, extracted
ISO 8601 timestamp of the sign-off.
- TranslationStatus.signedOffBy candidate, extracted
Who signed off. A human identifier (GitHub username, email, etc.).
- TranslationStatus.sourceHash candidate, extracted
SHA-256 hash of the source
.mdcontent at sign-off time. Used for staleness detection: if the current source hash differs, the translation is stale.- TranslationStatus.stale candidate, extracted
Whether the source content has changed since the last sign-off. Only meaningful when
officialis true.- TranslationStatusSchema.changedAt candidate, extracted
YYYY-MM-DD: the day the counts last changed.- TranslationStatusSchema.instances candidate, extracted
Every OTHER instance's declared
translation-sourcesdirectory, each measured on its own (issue #2228). Optional: an index written before the field existed is still a valid index.- TranslationStatusSchema.page candidate, extracted
The status page, site-relative, which the translation drawer links as "Full translation status" (bean
cat-tools-bbnd). Optional: an index written before the drawer existed has none.- TranslationStatusSchema.scope candidate, extracted
The translations directory counted, repository-relative.
U
- UnboundSidecarSchema.enumeration candidate, extracted
Which enumeration listed it.
- UnboundSidecarSchema.filename candidate, extracted
The sidecar file the enumeration named.
- UnboundSidecarSchema.reason candidate, extracted
Why no artefact matched: the strategies tried, in order.
- UnboundSidecarSchema.title candidate, extracted
Its title, as the enumeration gives it — usually the only human handle on what was missed.
- UndeterminedSchema.hash candidate, extracted
The input's blob at head, so a waiver of it is pinned to this version.
- UndeterminedSchema.scope candidate, extracted
all: the input can re-render anything (a site config, a shared include).- UniverseLevel.morphisms candidate, extracted
Universe for morphisms (typically
v).- UniverseLevel.objects candidate, extracted
Universe for objects (typically
u).- Unkinded.family candidate, extracted
The
$schemafamily; absent when it is the typology's kind-level validator.- UntaintedDispatch.adjudicator_sees candidate, extracted
Artefacts the ADJUDICATOR is given, alongside the checker's output.
- UntaintedDispatch.checker_sees candidate, extracted
Artefacts the CHECKER is given.
- UntaintedDispatch.checker_withheld candidate, extracted
Artefacts deliberately WITHHELD from the checker.
- UntaintedDispatch.drift candidate, extracted
What counts as a finding for this criterion, and what does not.
- UserScenarioBlock.personas candidate, extracted
Ids of the personas this scenario involves, at least one.
- UserStorySchema.role candidate, extracted
The role this story is told as.
- UserStorySchema.soThat candidate, extracted
Why, when it is not obvious from the want — "so that …".
- UserStorySchema.want candidate, extracted
What the reader wants to do, in their terms — "I want to …".
V
- ValidatorNodeSchema.id candidate, extracted
This validator's own name, as its file is named.
- ValidatorNodeSchema.pages candidate, extracted
The kind's own page renderer,
<harness>:<path>#<Export>, when the generic pages built from its schema are not enough (issue #2195: *"Generic + override"*). The export is aKindPages(scripts/gen-node-kind-pages.ts). Declared here, beside the schema, so the harness whose code knows the kind is the one that says how to show it.- ValidatorNodeSchema.rationale candidate, extracted
Why this validator is the one, when that is not obvious from the code.
- ValidatorNodeSchema.schema candidate, extracted
The Zod export,
<harness>:<path>#<Export>(or instance-relative<path>#<Export>).- ValidatorNodeSchema.validates candidate, extracted
What it validates: a graph typology, and the
$schemafamily within it when the kind has several.- ValidatorRef.exportName candidate, extracted
The exported symbol — required; see §"
module#Export".- ValidatorRef.instance candidate, extracted
The instance the module belongs to, by its DECLARED NAME, when that is not the instance resolving the reference:
folio-assistant-core:schemas/x.ts#X.- ValidatorRef.module candidate, extracted
Module path, relative to {@link ValidatorRef.instance}'s root, or to the resolving one.
- VectorAssemblySchema.background candidate, extracted
Of those, page-sized backgrounds that joined no component.
- VectorAssemblySchema.between candidate, extracted
Lines inside the region but in no component — the white space
a8wyfound real labels in.- VectorAssemblySchema.drawings candidate, extracted
Drawing operations on the page. Reported, never graded.
- VectorAssemblySchema.outside candidate, extracted
Lines outside the rendered region: furniture, and sometimes figure titles. The crop's stated limit.
- VectorAssemblySchema.unlabelledClosedShapes candidate, extracted
Closed shapes holding no label — arrowheads, bullets, swatches. Counted, not listed.
- VectorBoxSchema.parent candidate, extracted
The smallest other labelled box containing this one, or null.
- VectorConnectorSchema.from candidate, extracted
Box the stroke starts in, or null when it lands in none.
- VectorFigurePageSchema.captionCandidates candidate, extracted
Verbatim line-start
Fig./Figurematches on the page, table-of-contents entries removed. A lower bound, and it deliberately includes cross-references — plural, and never asserted to be the page's figure.- VectorFigurePageSchema.drawings candidate, extracted
Vector drawing operations MuPDF reports on the page. Reported, never graded.
- VectorFigurePageSchema.page candidate, extracted
1-based, matching
DocumentImage'sbasis.pageso the two join.- VectorFigurePageSchema.rasterImagesOnPage candidate, extracted
Raster images the page also places.
- VectorFigurePageSchema.rotation candidate, extracted
The page's rotation in degrees, as the PDF declares it.
- VectorFigureSchema.captionLabels candidate, extracted
Figure numbers from the page's caption candidates — candidates, not claims.
- VectorFigureSchema.file candidate, extracted
The render, relative to the entry; null when nothing was rendered, with the reason.
- VectorFigureSchema.id candidate, extracted
vfig-p034— one per qualifying page; a verdict keys on it.- VectorFigureSchema.region candidate, extracted
The rendered rectangle, visible frame.
- VectorFigureSchema.shows candidate, extracted
Which declared figures the render shows. Inspection only — see the module docblock.
- VectorFiguresSidecarSchema.figures candidate, extracted
nullis could-not-determine, never "draws no figures" — thedh4frule.- VectorLabelSchema.bbox candidate, extracted
The line's rectangle **in the visible frame** — rotation already applied.
- VectorLabelSchema.fonts candidate, extracted
Font names on the line's spans, deduplicated and sorted.
- VectorLabelSchema.intersectsDrawing candidate, extracted
Whether this block's rectangle intersects any vector drawing's.
- VectorLabelSchema.sizes candidate, extracted
Point sizes on the line's spans, deduplicated and sorted ascending.
- VectorLabelSchema.text candidate, extracted
The line's text: its spans concatenated, since a span break is a style change and not a word break.
- VectorLabelsSidecarSchema.pages candidate, extracted
nullis could-not-determine and is NOT an empty list.- VerdictIngestResult.malformed candidate, extracted
Said, not dropped, so the reviewer can fix it.
- VerificationEntry.fix_commit candidate, extracted
Reference to the commit that landed the fixes, if any.
- VerificationEntry.fixes_applied candidate, extracted
Number of citation-site fixes applied as part of this verification.
- VerificationEntry.human_adjudicated candidate, extracted
Separate human-adjudication provenance. Distinct from
verified_by: an entry can beverified_by: { kind: "agent", ... }and *also* carryhuman_adjudicatedif a human has subsequently reviewed the agent's claim. Status transitions likepaper-mismatchanduncited(which trigger reference removal / metadata rewrite) SHOULD require this field to be set.- VerificationEntry.id candidate, extracted
Reference id (matches a key in
references.ts).- VerificationEntry.local_pdf candidate, extracted
Path to a local PDF that was inspected, if any.
- VerificationEntry.note candidate, extracted
Free-text note for the verifier's reasoning.
- VerificationEntry.status candidate, extracted
Current verification status.
- VerificationEntry.verified_at candidate, extracted
ISO 8601 timestamp of the verification.
- VerificationEntry.verified_by candidate, extracted
Who / what produced the verification claim.
Statement of what this file is authoritative for.
- VerificationRoster._schema candidate, extracted
Self-describing schema string.
- VerificationRoster.entries candidate, extracted
The verification entries.
The layout criteria this railed page FAILS. Absent when it fails none.
The page's published path, e.g.
/cat-harness/library/<instance>/.Why, for anything other than
railed.Repo-relative source, so a finding names a file somebody can open.
Every page, sorted by published path, so the diff is stable.
The published docs root the audit walked, repo-relative.
- VisualisationSchema.alias candidate, extracted
Opt-in top-level route alias:
<base>/<alias>/(bean folio-assistant-t4xb). Canonical route is always<base>/<harness>/<visualizer>/. When true, uses the visualizer's name as the alias; when string, uses the specified alias. Refused if it collides with another alias, a harness route, or a reserved site route.Whether this tile starts out of frame. Absent means shown.
- VisualisationSchema.icon candidate, extracted
WHICH GLYPH the tile wears, by NAME. Absent falls back to the generic node-graph glyph every tile shared before this field existed.
- VisualisationSchema.publish candidate, extracted
WHERE this visualisation may be published. Absent means everywhere, which is what every visualisation did before this field existed.
- VisualisationSchema.ref candidate, extracted
The page that renders it, **relative to the REPOSITORY root** — see {@link SubgraphCoverageSchema.visualiser}.
- VisualisationSchema.surfaces candidate, extracted
Where its tile appears. Absent means EVERY surface in {@link TILE_SURFACES}.
- VisualisationSchema.theme candidate, extracted
The tile's theme, by reference (#1168 B8). Absent means the directory's, then the instance's.
- VisualisationSchema.title candidate, extracted
What a tile calls it. Absent, the tile takes the display name of the directory's first graph typology (
kindTitle, viascripts/lib/nav-label.ts), which is what every other surface calls that page. Declare one only when the page is not simply "the kind's viewer". Never add the harness as a suffix ("Docs — cat-harness"): the qualifier is appended by the surface (beanob3mfinding 6, "One name everywhere").- VisualisationSchema.writer candidate, extracted
The GENERATOR that writes this page: repo-relative script files, run with
bun run <file>— the same field, with the same meaning, as a directory'swriter. Bean0b8c(#2230).- VisualiserRouteParts.asset candidate, extracted
A path below the (sub-graph's) view,
/-separated, no leading slash.- VisualiserRouteParts.harness candidate, extracted
The declaring harness's
name.- VisualiserRouteParts.locale candidate, extracted
The locale the page is rendered in (
en), the route's first segment. Absent for the unlocalised route the t4xb rule keeps working.- VisualiserRouteParts.subgraph candidate, extracted
A sub-graph of what the visualiser covers — absent for the full KG. One segment, or a
/-separated path of segments when the visualiser declares its sub-graphs nested (auto-docs'index/skills/core-skills, owner 2026-10-03: *"auto-docs is one declared subgraph, with declared sub-sub-graphs per writer"*).- VisualiserRouteParts.visualiser candidate, extracted
The visualiser's declared
id.- VizCamera.phi candidate, extracted
Polar angle in radians.
- VizCamera.r candidate, extracted
Distance from target point.
- VizCamera.target candidate, extracted
Camera look-at target (world coords). Defaults to origin.
- VizCamera.theta candidate, extracted
Azimuthal angle in radians.
- VizCone.gamma candidate, extracted
Pitch scale (γ in ψᵢ(s)=2πi/3+q·2πs·Q·γ).
- VizCone.N_max candidate, extracted
End of tube segment in Planck lengths.
- VizCone.N_min candidate, extracted
Start of tube segment in Planck lengths.
- VizCone.sigma candidate, extracted
Cone half-angle (σ in ρ(s)=s·ℓ_P·σ·Ξ_q).
- VizState.blowup candidate, extracted
When true, renders the exceptional divisor E_p as an opaque ℂP¹ surface.
- VizState.brane_point candidate, extracted
Pre-normalisation (x,y,z) of the selected point p on the brane S³.
- VizState.camera candidate, extracted
Spherical-coordinate camera orbit.
- VizState.cone candidate, extracted
Parameters for cone-tube mode.
- VizState.mode candidate, extracted
Rendering mode.
- VizState.q candidate, extracted
Torsion / chirality parameter (substrate parameter). q=0 is classical.
- VizState.tubes candidate, extracted
Tube / tile visibility and opacity.
- VizState.visibility candidate, extracted
Scene element visibility flags.
- VizVisibility.axes candidate, extracted
Show xyz axes helper.
- VizVisibility.axispts candidate, extracted
Show colour tubes at ±axis poles.
- VizVisibility.brane candidate, extracted
Show brane sphere S³.
- VizVisibility.stars candidate, extracted
Show background star field.
- VocabMappingSchema.id candidate, extracted
This table's id. A ConceptMap's own resource
idis metadata, kept inmetadata.id.- VocabMappingSchema.metadata candidate, extracted
Every other ConceptMap field (contact, jurisdiction, meta, text, identifier, the resource id, …), verbatim.
- VocabMappingSchema.property candidate, extracted
R5 map-level
propertyandadditionalAttributedefinitions, verbatim.- VocabularyWrongSchema.concept candidate, extracted
The concept the label reached, where it reached one.
- VocabularyWrongSchema.domain candidate, extracted
The domain it does not serve.
- VocabularyWrongSchema.vocabulary candidate, extracted
The vocabulary consulted, as the check's scope names it (
platform,who-smart-base@v1.0.0, …).- VoiceApplicabilitySchema.processes candidate, extracted
BPMN process ids, as
<bpmn:process id>spells them —Process_CRDM_Requirements, not a filename (#1168, B8: the element id, ascalledElementuses it). Resolved against the loaded diagrams byprocess-refs.test.ts, so a typo is a dangling reference rather than a voice that quietly never activates.- VoiceApplicabilitySchema.roles candidate, extracted
The roles this voice addresses — the reader it is written FOR.
- VoiceApplicabilitySchema.scenarios candidate, extracted
User scenarios or requirement ids this voice serves.
- VoiceConfigSchema.active candidate, extracted
Voice ids this folio activates. ABSENT OR EMPTY MEANS NONE — see the module comment on why the default is the empty set rather than everything.
- VoiceInstructionsSchema.file candidate, extracted
The skill body, as a sibling file —
SKILL.mdby convention.- VoiceInstructionsSchema.summary candidate, extracted
What an actor is expected to DO with this voice, one line, for a skill index that lists it beside every other skill.
- VoicePatternSchema.message candidate, extracted
What a match MEANS. Shown as the finding's evidence.
- VoiceProfileSchema.$schema candidate, extracted
EITHER tag, because a voice skill IS a profile plus the skill half.
- VoiceProfileSchema.activeIn candidate, extracted
When this voice is in force, by process, lane or scenario. Absent means everywhere. See {@link VoiceApplicabilitySchema}.
- VoiceProfileSchema.appliesTo candidate, extracted
What this voice audits — block kinds or artefact kinds. Absent means every kind the folio has.
- VoiceProfileSchema.extends candidate, extracted
The voice this one OVERRIDES — a shared base it adds to and narrows.
- VoiceProfileSchema.overlaySeverity candidate, extracted
How severe a finding against THIS VOICE AS A WHOLE is, for the overlay criterion derived from it.
- VoiceProfileSchema.provenance candidate, extracted
Where this voice's rules come from, epistemically. See {@link VOICE_PROVENANCE} for why it is required and has no default.
- VoiceProfileSchema.sources candidate, extracted
The ingested documents this voice is derived from.
- VoiceProfileSchema.superseded candidate, extracted
Set when the source has been replaced. See {@link VoiceSupersessionSchema}.
- VoiceProvenanceFlag.ruleIds candidate, extracted
Rule ids that cite a node of this instance's own graph.
- VoiceRefSchema.instance candidate, extracted
The instance holding it, as
owner/repo(bean6rmv). Absent means this one.- VoiceRuleSchema.judgementOnly candidate, extracted
True when no lexical check can decide this rule and a reviewer must. Recorded rather than inferred from the absence of
patterns, because "nobody has written the pattern yet" and "no pattern could work" are different states and only the second is a finished rule.- VoiceRuleSchema.title candidate, extracted
Required here even though
kgNodeLabelShapemakes it optional: an unnamed rule cannot be reported.- VoiceRuleSourceSchema.instance candidate, extracted
The instance holding the corpus, as its planned
owner/repo(bean6rmv; it was the declared name until then), when it is not this one. Absent means this instance, so every existing citation keeps its meaning unchanged.- VoiceRuleSourceSchema.libraryId candidate, extracted
doc_idunderlibrary/, e.g.who-pub-tps-931.- VoiceRuleSourceSchema.pages candidate, extracted
Page range in the source document, e.g. "14" or "131-132".
- VoiceRuleSourceSchema.quote candidate, extracted
The passage the rule was read from. Required: a citation with no quote cannot be checked without re-reading the source, which is the cost this field exists to remove.
- VoiceRuleSourceSchema.sectionId candidate, extracted
section_idwithin that document, e.g.page-014orsec-180-106-….- VoiceScope.activeIn candidate, extracted
Process, lane and scenario. Absent means everywhere.
- VoiceScope.appliesTo candidate, extracted
Block kinds. Absent means every kind the folio has.
- VoiceSkillRuleSchema.commonError candidate, extracted
Why the instinct is wrong, in one sentence, for a drafter who has not read the source.
- VoiceSkillRuleSchema.counterintuitive candidate, extracted
The instinct is wrong, so a drafter must know this rule BEFORE writing.
- VoiceSkillSchema.overlaySeverity candidate, extracted
How severe a finding against this voice AS A WHOLE is, for the overlay criterion derived from it.
- VoiceSkillSchema.owner candidate, extracted
The instance that DERIVED this voice, by declared name.
- VoiceSkillSchema.sources candidate, extracted
The ingested documents this voice is derived from.
- VoiceSupersessionSchema.by candidate, extracted
What replaced it, in the source's own words.
- VoiceSupersessionSchema.on candidate, extracted
ISO date the replacement took effect, as the source states it.
- VoiceSupersessionSchema.quote candidate, extracted
Where that is stated — a quote, so a reader need not re-open the PDF.
- VoiceSupersessionSchema.successor candidate, extracted
The voice reading the successor document, when one has been ingested.
- VoiceTerminologySchema.context candidate, extracted
When the pair applies only in a particular context.
W
- WaiverNodeSchema.$schema candidate, extracted
Tag identifying the file's contract, per the directory conventions.
- WaiverNodeSchema.bounds candidate, extracted
Anything the grant carried that the gate needs to check against — a swarm's agent count and model level, the issues an
issue-closewaiver names, the artefact class adeletionwaiver is limited to.- WaiverNodeSchema.expires candidate, extracted
When it stops applying (ISO 8601). Required — see the module note.
- WaiverNodeSchema.gate candidate, extracted
Which gate is waived. Never "everything" — see {@link WAIVABLE_GATES}.
- WaiverNodeSchema.grantedAt candidate, extracted
When it was granted (ISO 8601).
- WaiverNodeSchema.grantedBy candidate, extracted
The person granting it, by the identity this repository already knows them by.
- WaiverNodeSchema.id candidate, extracted
Node id — the file's basename.
- WaiverNodeSchema.quote candidate, extracted
The grantor's **own words**, verbatim.
- WaiverNodeSchema.scope candidate, extracted
How far it reaches.
- WebPage.documents candidate, extracted
What this page documents: graph typologies (
library), or one directory as<instance>/<id>where the kind is too general. Emitted into the page's front matter, where the coverage check reads it — the page names the directory, and the directory names no page (#1168 B7c).- WebPage.heading candidate, extracted
The page's
# heading, when it differs from the nav title. Two guides do differ today — the nav says "Writing a document" while the page opens "Writing a document with folio-assistant" — so this is a real distinction in the corpus, not a hypothetical. Omit it and the title is used for both.just-the-docs nav front matter.
- WebPage.nodes candidate, extracted
Ordered children.
- WebPage.slug candidate, extracted
URL stem.
document-ingestionpublishes todocs/guides/document-ingestion.mdand thus to<baseurl>/guides/document-ingestion.html, which is what existing inbound links already use.- WebPageAsset.alt candidate, extracted
Long description for the embed's
alt. Not optional in practice: these diagrams carry the argument of the page around them, and a reader who cannot see the SVG gets only this.- WebPageAsset.rendered candidate, extracted
The rendered artefact the page embeds, as a path relative to the published page. GENERATED — for
kind: "bpmn"this isrender-bpmn.tsoutput and carries its own "never hand-edit" rule, which is exactly why it is kept separate fromsourcerather than inferred from it.- WebPageAsset.source candidate, extracted
The asset's SOURCE, repo-root-relative. This is the edit target and the thing a human changes.
- WebPageAsset.sourceLinks candidate, extracted
The link line under the figure. TWO shapes exist in the corpus today and this models both rather than normalising one into the other:
- WebPageNode.asset candidate, extracted
An asset this node presents, with its own edit target.
- WebPageNode.block candidate, extracted
Root name of a block
.ts+.mdpair in the page directory, exactly asSection.blocksnames them. The.mdis this node's narrative and is the edit target for a narrative node. When the node carries an asset this is the prose AFTER it.- WebPageNode.id candidate, extracted
Stable identity. Becomes the heading anchor (pinned, not derived from the title), the local part of the node's
@id, and the fragment that inbound links — including the BPMN subprocess links derived fromasset.source— resolve against.- WebPageNode.lead candidate, extracted
Narrative that INTRODUCES the asset, emitted before it.
- WebPageNode.level candidate, extracted
Heading depth under the page's
# title. Default 2.- WebPageNode.title candidate, extracted
Heading text. Free to change without breaking a link.
- WithheldEntrySchema.gates candidate, extracted
The gates that refused, STRUCTURED — issue #1794. Optional because a generator may know only the sentence; when present, a reader shows "copyright not granted" from the gate itself rather than parsing {@link reason}, which is prose and free to change shape.
- WithheldEntrySchema.path candidate, extracted
Relative to the directory holding the file; a directory takes its whole subtree.
- WithheldEntrySchema.reason candidate, extracted
Why — which gate, which verdict. A withholding with no reason cannot be re-checked.
- WithheldEntrySchema.record candidate, extracted
Where a reader can see what the work IS without being shown the work — the catalogue record.
pageis a SITE-ROOT path (leading/) to a published record page;uriis the upstream identifier (a Handle IRI, say). Either may be absent; a viewer linkspagefirst,urisecond, and nothing at all rather than a guess.- WorkflowDefinition.bindings candidate, extracted
Skill assignments per stage.
- WorkflowDefinition.description candidate, extracted
Description of this workflow.
- WorkflowDefinition.id candidate, extracted
Unique workflow identifier.
- WorkflowDefinition.name candidate, extracted
Human-readable name.
- WorkflowDefinition.stages candidate, extracted
Ordered lifecycle stages used by this workflow.
- WorkflowDefinition.transitions candidate, extracted
Allowed transitions between stages.
Whether an AUTHOR wrote this route's content, as opposed to a generator regenerating it.
- WriteRequestSchema.requestedBy candidate, extracted
Who asked — a workflow file, a session URL, or an actor id.
- WriteRequestSchema.route candidate, extracted
The route this write owns —
/for the published site,STAGING/<slug>/for a preview. Composition is per route: two different routes compose by union, two of the same route take the newer.- WriteResult.written candidate, extracted
Paths written (store) or the legacy path.