Skip to content

GaZmagik/iso-24495 / skills/iso-24495-4

iso-24495-4

Verified

Provisional task skill for organisational plain language implementation (based on ISO/CD 24495-4, committee draft). Activates for plain language gap analysis, policy drafting, review workflow design, and organisational readiness for the future published standard. Does not activate for ordinary writing, rewriting, or reviewing of individual documents.

FileKindSize
SKILL.mdskill md5.2 KB
agents/openai.yamlconfig300 B
assets/gap-report-template.mdmarkdown1.3 KB
references/evidence-map.mdmarkdown1.4 KB
references/interview-guide.mdmarkdown1.4 KB
references/maturity-model.mdmarkdown2.7 KB
scripts/audit-corpus-cli.tsscript109 B
scripts/audit-corpus.tsscript50.8 KB
scripts/audit-evidence-cli.tsscript111 B
scripts/audit-evidence.tsscript3.1 KB
scripts/generate-report-cli.tsscript112 B
scripts/generate-report.tsscript5.0 KB
scripts/lib/lexicon.tsscript88.8 KB
scripts/lib/parse.tsscript49.8 KB
scripts/lib/types.tsscript1.6 KB
scripts/score-maturity-cli.tsscript111 B
scripts/score-maturity.tsscript2.3 KB
tests/audit-corpus.test.tsscript24.6 KB
tests/audit-evidence.test.tsscript1.4 KB
tests/behaviour-contract.test.tsscript111.1 KB
tests/cli.test.tsscript26.0 KB
tests/fixtures/answers.sample.jsonconfig880 B
tests/fixtures/corpus/average-heavy.mdmarkdown1.6 KB
tests/fixtures/corpus/code-fenced.mdmarkdown320 B
tests/fixtures/corpus/dense-paragraphs.mdmarkdown362 B
tests/fixtures/corpus/good-policy.mdmarkdown201 B
tests/fixtures/corpus/headings-broken.mdmarkdown115 B
tests/fixtures/corpus/legalese-sample.mdmarkdown226 B
tests/fixtures/corpus/long-sentences.mdmarkdown656 B
tests/fixtures/difficult/every-rule.mdmarkdown2.1 KB
tests/fixtures/difficult/us-foia-statute.mdmarkdown770 B
tests/fixtures/known-good/SOURCES.mdmarkdown3.7 KB
tests/fixtures/known-good/consumer-rights.mdmarkdown174 B
tests/fixtures/known-good/govuk-universal-credit.mdmarkdown1.1 KB
tests/fixtures/known-good/international-english.mdmarkdown206 B
tests/fixtures/known-good/names-and-measures.mdmarkdown161 B
tests/fixtures/known-good/public-service.mdmarkdown225 B
tests/fixtures/known-good/technical-guide.mdmarkdown208 B
tests/fixtures/known-good/warning.mdmarkdown127 B
tests/fixtures/reference-blocks.tsscript67.0 KB
tests/fixtures/repo-level0/docs/notes.mdmarkdown81 B
tests/fixtures/repo-level2/.github/PULL_REQUEST_TEMPLATE.mdmarkdown69 B
tests/fixtures/repo-level2/.github/workflows/text-lint.ymlconfig179 B
tests/fixtures/repo-level2/docs/plain-language-policy.mdmarkdown109 B
tests/fixtures/repo-level2/glossary.mdmarkdown89 B
tests/fixtures/repo-level2/training/introduction.mdmarkdown96 B
tests/generate-report.test.tsscript5.4 KB
tests/reference/README.mdmarkdown1.6 KB
tests/reference/build-fixture.tsscript8.1 KB
tests/reference/shapes.tsscript6.4 KB
tests/robustness.test.tsscript2.4 KB
tests/score-maturity.test.tsscript1.4 KB
tests/text-audit.test.tsscript11.0 KB

tests/reference/README.md

View on GitHub
# Comparing this parser with the CommonMark reference

`../fixtures/reference-blocks.ts` records how this engine reads a corpus of documents, and how many of them the CommonMark reference implementation reads the same way. `build-fixture.ts` produced it, so the claim can be checked rather than believed.

The reference is not a dependency of this project. Install it outside the repository and run one command:

```sh
mkdir /tmp/cmark && cd /tmp/cmark
bun add commonmark
bun /path/to/skills/iso-24495-4/tests/reference/build-fixture.ts
```

It prints the counts and rewrites the fixture in place. Nothing else is needed, and no intermediate file has to exist first.

`shapes.ts` holds the corpus, in two parts. One is a matrix: every container prefix the engine claims to understand, applied to every leaf block it claims to find, and the transitions between them. The other is built from a grammar of lines by a seeded generator, so the corpus holds shapes nobody chose. The seed is fixed, so the corpus is the same every time.

**A difference the engine cannot explain stops the run.** `reasonFor` decides whether this engine reads a document differently on purpose. It decides from what the document contains, not from a word in its name. An earlier version classified by name, which would have let an unrelated defect wear an unrelated excuse.

One adjustment makes the comparison meaningful. The reference renders inline markup, while this engine keeps the source, because its rules read link syntax and code spans. The comparison flattens ours. Without that, 153 documents looked like disagreements and every one was the comparison's fault.
brainstormingobra/superpowersYou MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).Stars: 322kUpdated Claude CodeCodexVerifiedInstall
brand-guidelinesanthropics/skillsApplies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
frontend-designanthropics/skillsGuidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
mcp-builderanthropics/skillsGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).Stars: 198kUpdated Claude CodeCodexVerifiedInstall