L1 guideline → recommendation candidates (L1 3.0)

Find every normative sentence of an ingested library/ entry (folio-assistant-core’s l1-coverage markers), flag uncertain readings, and propose the publication’s identity. Given an agent’s l1-recommendation-curation/v1, refuse a non-verbatim statement or an unknown location, and write the recommendations and remarks as an L1 3.0 graph (checked with L1/schemas/validate.ts) with coverage and proposed, unsigned exclusions.

Skills it satisfies

A skill states the capability; this Tool is one way to exercise it. Each skill’s declared input and output contract follows, where it has one.

recommendation-candidates

Turn one ingested L1 source into L1 3.0 recommendation nodes: list every normative sentence as a candidate, let the agent decide each in a curation file, then refuse a non-verbatim statement or unknown location and write the graph, coverage and proposed exclusions. Step 1 of recommendation-extraction, run once per included source; the l1-recommendation-candidates Tool satisfies it.

Source: smart-base/L1/skills/smart-l1/recommendation-candidates.md.

input [1..1]output [1..1]recommendation-candidates«skill»recommendation-candidatesinput : RecommendationCandidatesInput [1..1]output : RecommendationCandidatesOutput [1..1]RecommendationCandidatesInput«json-schema»RecommendationCandidatesInputentry : string [1..1]curation : string [0..1]out : string [0..1]RecommendationCandidatesOutput«json-schema»RecommendationCandidatesOutputcandidates : string [1..1]graph : string [0..1]exclusions : string [0..1]warnings : string [0..1]runContext : string [1..1]

input — smart-base/L1/schemas/skills/recommendation-candidates/input.schema.json

fieldtyperequireddescription
entrystringyesAn ingested library// with page-tagged sections/*.md and structure.json.
curationstringThe agent’s decision per candidate location (l1-recommendation-curation/v1). Absent: only the work list is written.
outstringWhere the outputs are written.

output — smart-base/L1/schemas/skills/recommendation-candidates/output.schema.json

fieldtyperequireddescription
candidatesstringyesrecommendation-candidates.json: every normative sentence, its hints and its decision.
graphstringl1.json: the L1 3.0 publication, recommendation and remark nodes, every one inferred. Present once a curation is given.
exclusionsstringexclusions.proposed.json, in l1-coverage’s shape; none counts until a person signs it.
warningsstringextraction-warnings.md: undecided candidates, unsigned exclusions, statements from a layout fragment.
runContextstringyesrun-context.json, cat-harness folio-run-context/v1.

Inputs and outputs

Each port names a shared type by IRI, so two Tools for one skill can be checked for interchangeability.

porttyperequireddescriptionconstraint
inentryRepoPathyesAn ingested library/<id>/ with page-tagged sections/*.md and structure.json.string · pattern ^[A-Za-z0-9._][A-Za-z0-9._/-]*$ · A path relative to the repository root. No .. segments.
incurationRepoPathThe agent’s decisions per candidate location (l1-recommendation-curation/v1). Absent: only the work list is written.string · pattern ^[A-Za-z0-9._][A-Za-z0-9._/-]*$ · A path relative to the repository root. No .. segments.
outcandidatesRepoPathrecommendation-candidates.json — every normative sentence, its hints and its decision; cites the run context.string · pattern ^[A-Za-z0-9._][A-Za-z0-9._/-]*$ · A path relative to the repository root. No .. segments.
outgraphRepoPathl1.json — L1 3.0 publication, recommendation and remark nodes, every one inferred.string · pattern ^[A-Za-z0-9._][A-Za-z0-9._/-]*$ · A path relative to the repository root. No .. segments.
outexclusionsRepoPathexclusions.proposed.json, in l1-coverage’s shape; none counts until a person signs it.string · pattern ^[A-Za-z0-9._][A-Za-z0-9._/-]*$ · A path relative to the repository root. No .. segments.
outwarningsRepoPathextraction-warnings.md — undecided candidates, unsigned exclusions, statements taken from a layout fragment.string · pattern ^[A-Za-z0-9._][A-Za-z0-9._/-]*$ · A path relative to the repository root. No .. segments.
outrunContextRepoPathrun-context.json, cat-harness folio-run-context/v1.string · pattern ^[A-Za-z0-9._][A-Za-z0-9._/-]*$ · A path relative to the repository root. No .. segments.

Processes and tasks

No process step names a skill this Tool satisfies.

How it is deployed and reached

  • Command line. From the repository root: bun run smart-base/L1/scripts/l1-recommendation-candidates.ts --entry <library entry> [--curation <curation.yaml>] --out <dir> [--check].
  • GitHub Actions. No workflow in this checkout runs its command.

As an API: its operation in the OpenAPI document and the MCP tool list of its layer.

installnone
runtimebun
networkno
when to choose it
whenA WHO guideline, position paper or handbook has been ingested and its recommendations are to become L1 3.0 nodes — especially one with no Recommendation N: labels, where extract-smart-kg-l1 finds none.
limitsFinds candidates by marker words only: a recommendation without should/must/recommend… (a table step, an imperative) is invisible to it. The splitter reads a bulleted list or table row as one sentence; the curation’s statement override narrows it. It decides nothing: with no curation every candidate is undecided.
costSeconds, no network.

What it is made of

Linked data

The Tool is published as its own JSON-LD document at its IRI, http://smart.who.int/base/tools/l1-recommendation-candidates.jsonld, partOf its layer’s document. The node below is the definition this page was drawn from, and is embedded in the page as application/ld+json.

{
  "@context": "http://smart.who.int/base/smart-base.context.jsonld",
  "@id": "http://smart.who.int/base/tools/l1-recommendation-candidates.jsonld",
  "@type": "https://litlfred.github.io/cat-harness/v0/ns/Tool",
  "partOf": "http://smart.who.int/base/smart-base.jsonld",
  "name": "l1-recommendation-candidates",
  "title": "L1 guideline → recommendation candidates (L1 3.0)",
  "description": "Find every normative sentence of an ingested `library/` entry (folio-assistant-core's `l1-coverage` markers), flag uncertain readings, and propose the publication's identity. Given an agent's `l1-recommendation-curation/v1`, refuse a non-verbatim statement or an unknown location, and write the recommendations and remarks as an L1 3.0 graph (checked with `L1/schemas/validate.ts`) with coverage and proposed, unsigned exclusions.",
  "install": {
    "none": true
  },
  "invoke": {
    "shell": "bun run smart-base/L1/scripts/l1-recommendation-candidates.ts --entry <library entry> [--curation <curation.yaml>] --out <dir> [--check]"
  },
  "io": {
    "inputs": [
      {
        "name": "entry",
        "schema": "https://litlfred.github.io/folio-assistant/cat-harness/tool-types.schema.json#/$defs/RepoPath",
        "description": "An ingested `library/<id>/` with page-tagged `sections/*.md` and `structure.json`.",
        "required": true
      },
      {
        "name": "curation",
        "schema": "https://litlfred.github.io/folio-assistant/cat-harness/tool-types.schema.json#/$defs/RepoPath",
        "description": "The agent's decisions per candidate location (`l1-recommendation-curation/v1`). Absent: only the work list is written.",
        "required": false
      }
    ],
    "outputs": [
      {
        "name": "candidates",
        "schema": "https://litlfred.github.io/folio-assistant/cat-harness/tool-types.schema.json#/$defs/RepoPath",
        "description": "`recommendation-candidates.json` — every normative sentence, its hints and its decision; cites the run context."
      },
      {
        "name": "graph",
        "schema": "https://litlfred.github.io/folio-assistant/cat-harness/tool-types.schema.json#/$defs/RepoPath",
        "description": "`l1.json` — L1 3.0 publication, recommendation and remark nodes, every one `inferred`."
      },
      {
        "name": "exclusions",
        "schema": "https://litlfred.github.io/folio-assistant/cat-harness/tool-types.schema.json#/$defs/RepoPath",
        "description": "`exclusions.proposed.json`, in `l1-coverage`'s shape; none counts until a person signs it."
      },
      {
        "name": "warnings",
        "schema": "https://litlfred.github.io/folio-assistant/cat-harness/tool-types.schema.json#/$defs/RepoPath",
        "description": "`extraction-warnings.md` — undecided candidates, unsigned exclusions, statements taken from a layout fragment."
      },
      {
        "name": "runContext",
        "schema": "https://litlfred.github.io/folio-assistant/cat-harness/tool-types.schema.json#/$defs/RepoPath",
        "description": "`run-context.json`, cat-harness `folio-run-context/v1`."
      }
    ]
  },
  "satisfies": [
    "recommendation-candidates"
  ],
  "requirements": {
    "runtime": [
      "bun"
    ],
    "network": false
  },
  "selection": {
    "when": "A WHO guideline, position paper or handbook has been ingested and its recommendations are to become L1 3.0 nodes — especially one with no `Recommendation N:` labels, where `extract-smart-kg-l1` finds none.",
    "limits": "Finds candidates by marker words only: a recommendation without `should`/`must`/`recommend…` (a table step, an imperative) is invisible to it. The splitter reads a bulleted list or table row as one sentence; the curation's `statement` override narrows it. It decides nothing: with no curation every candidate is `undecided`.",
    "cost": "Seconds, no network."
  }
}

folio-assistant — code under the Apache License 2.0, documentation under CC BY 3.0. Generated schema and API references are rebuilt from source on every push to main.

deployed dev