メニュー
Use when files under docs/ may be stale. Discovers commits since the last docs sweep and reconciles each docs/*.md file with its sources of truth (component CSS, schemas, search index, design tokens).
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when the user wants to add, update, or remove entries in the CV (src/data/cv.json), or wants recommendations on what to add, remove, or emphasize. Always validates the result against schemas/cv.schema.json before finishing.
Use when the Visual workflow (.github/workflows/visual.yml) is failing on `main` or on a PR. Reproduces the failure, identifies which commit broke each snapshot, decides whether the diff is intentional, re-records baselines if so, and commits the new pixels.
Use when the user wants to clean up the open Dependabot PR queue. Walks the open Dependabot PRs newest-first, merges the green ones, then rebases and fixes the CI-failing ones, pushes, waits for CI, and merges.
Use to bring the whole repo back in sync after a batch of changes. Routes to every update-* skill in deterministic order and aggregates the resulting diff.
Use when this repo may have drifted out of conformance with OSS_SPEC.md. Runs the bash mirror of the oss-spec validator (no Rust toolchain required), walks the violations, and fixes each one until the validator reports zero violations.
Use when the user wants to refresh the `tagline` and/or `description` of one or more companies in `src/data/cv.json` using their declared `sourceUrls`. Fetches each source, synthesizes a tight bilingual tagline + description pair, propagates the change into any matching `experience[].printDescription` (and assignment `printDescription` when the assignment's `clientId` resolves to a company being refreshed), and validates the result.
| name | update-docs |
| description | Use when files under docs/ may be stale. Discovers commits since the last docs sweep and reconciles each docs/*.md file with its sources of truth (component CSS, schemas, search index, design tokens). |
Bring the prose under docs/ back in sync with the code, schemas, and
data files those documents describe. docs/ currently ships:
docs/DESIGN.md — visual design system (tokens, layout, motion).docs/SEARCH.md — search index format, ranker behaviour, alias
conventions.docs/getting-started.md — contributor onboarding (prereqs, scripts,
local-override conventions).Each doc has a different source of truth. This skill walks them in order, asks "did the source change since the baseline?", and rewrites only the sections that drifted.
The sibling file .last-updated holds the git commit hash of the last
successful run. An empty file means "never run" — use the initial
commit as the baseline.
Determine what has changed since the baseline:
BASE=$(cat .agent/skills/update-docs/.last-updated)
HEAD=$(git rev-parse HEAD)
BASE=${BASE:-$(git rev-list --max-parents=0 HEAD)}
git log --oneline "$BASE..$HEAD"
git diff --name-only "$BASE..$HEAD"
Read every docs/*.md file end-to-end before editing. The files are
small enough that the round-trip is cheap and ensures the rewrite
matches the existing voice and structure.
| Source of truth | Docs file / section |
|---|---|
src/styles/tokens.css | docs/DESIGN.md → Tokens |
src/styles/hero.css / focus.css / experience.css / projects.css / print.css | docs/DESIGN.md → per-section style chapter |
src/styles/timeline-vis.css | docs/DESIGN.md → Timeline |
src/components/Hero.tsx / Focus.tsx / Experience.tsx / Projects.tsx | docs/DESIGN.md → section component notes |
Any new src/components/*.tsx section | docs/DESIGN.md → add a section chapter |
schemas/search-index.schema.json | docs/SEARCH.md → Index shape |
scripts/generate-search-index.mjs | docs/SEARCH.md → Index generation |
src/utils/search.ts | docs/SEARCH.md → Ranker behaviour, scoring weights |
src/components/SearchModal.tsx + src/styles/search.css | docs/SEARCH.md → UI behaviour |
aliases arrays under src/data/cv/*.json | docs/SEARCH.md → Aliases section |
package.json scripts / Makefile targets | docs/getting-started.md → Scripts table |
.nvmrc, .github/workflows/ci.yml Node version | docs/getting-started.md → Prerequisites |
src/data/cv.local.example.json / src/data/facts.local.example.json | docs/getting-started.md → Local overrides |
When a new doc is added under docs/, append a row that maps its
upstream sources, and update AGENTS.md § "Where new code goes" to
point to it.
docs/DESIGN.md: tokens listed match src/styles/tokens.css;
each section chapter exists for every component currently
mounted in App.tsx; the per-section style files referenced
still exist under src/styles/.docs/SEARCH.md: shape matches
schemas/search-index.schema.json; ranker weights and field
lists match src/utils/search.ts; alias conventions match
what scripts/generate-search-index.mjs actually consumes.docs/getting-started.md: scripts/targets table matches
package.json + Makefile; Node version matches .nvmrc;
local-override notes reference the live *.example.json
filenames.README.md and AGENTS.md to specific docs
sections still resolve (anchors haven't been renamed).docs/*.md end-to-end and confirm every claim
against the source(s) it references.make fmt-check so Prettier-managed prose stays consistent.make build if a doc cross-links from runtime code (it
shouldn't, but the type-check pass is cheap insurance).When a new source-of-truth file starts driving a docs section, add a row to the mapping table. When a doc file is renamed or split, update this skill alongside the rename so the registry stays accurate. Commit skill edits in the same PR as the docs edits they unblock.