Skip to content

Fix stale shipped kit docs and counts, and bump kit versions - #90

Draft
AtlasScope wants to merge 1 commit into
mainfrom
claude/kit-plugin-docs-bumps
Draft

AtlasScope wants to merge 1 commit into
mainfrom
claude/kit-plugin-docs-bumps

Conversation

@AtlasScope

Copy link
Copy Markdown
Owner

Split out from #88 so shipped plugin doc/skill fixes carry their required version bumps (design-kit 1.4.0 -> 1.4.1, engineering-kit 1.6.0 -> 1.6.1, experience-architecture 0.4.0 -> 0.4.1) without mixing non-doc manifest/library changes into the root docs refresh PR.

What changed

  • Plugin READMEs:
    • kits/design-kit/plugins/design-kit/README.md: updated "two choices" to "three choices" (start_mode, ask_before_edits, router_log), clarified working directories for scripts/build-claude-ai-zips.sh and kits/design-kit/scripts/impeccable-helpers.json, and aligned "borrowed skills" phrasing.
    • kits/engineering-kit/plugins/engineering-kit/README.md: noted library/library.json as the catalog source, updated /engineering-setup to describe both install choices (start_mode and router_log), and noted that node library/check.mjs checks the /database reference.
    • kits/experience-architecture/plugins/experience-architecture/README.md: fixed the stale --plugin-dir path to <clone>/kits/experience-architecture/plugins/experience-architecture.
  • Skill files touched (stale counts/singulars only; frontmatter keys, names, file paths, and structure unchanged):
    • kits/design-kit/claude-ai/SKILL.md: updated "about thirty specialist skills, twenty-two recipes" to "36 specialist skills, 23 recipes".
    • kits/engineering-kit/plugins/engineering-kit/skills/engineer/SKILL.md: updated "50 investigations, 10 thinking prompts and two dozen skills" to "56 investigations, 12 thinking prompts and 30 skills" (matching references/investigations.md and references/start-here.md), and updated line 17 from "all 50" to "all 56".
    • kits/engineering-kit/plugins/engineering-kit/skills/engineering-setup/SKILL.md: updated "this question" / "the same choice" to "these questions" / "the same choices" to match its two options.
  • Documented version bumps:
    • Raised design-kit to 1.4.1, engineering-kit to 1.6.1, and experience-architecture to 0.4.1 in .claude-plugin/marketplace.json and both .claude-plugin/plugin.json and .codex-plugin/plugin.json manifests for each kit.
    • Updated engineering-kit and experience-architecture version badges in library/library.json and rebuilt library/dist/ with node library/build.mjs.

Verification

  • node scripts/check-public.mjs: passed (No personal references in 634 shipped files.)
  • node scripts/check-catalog.mjs: passed (Catalogs OK: 5 plugins, 64 skills, both catalogs agree.)
  • node library/check.mjs: passed (Library OK: 254 entries, 18 database prompts, 5 database workflows.)
  • node tests/routing/check.mjs: passed (Routing cases OK: 59 cases (48 stable, 11 held out).)
  • (cd kits/experience-architecture && npm test): passed (25 passed, 0 failed, 4 skipped because Northwind wasn't running on :3100)
  • (cd site && npm ci --prefer-offline --no-audit --no-fund && npm run content && npm run typecheck && npm test && npx next build): passed (content: 254 prompts, 32 areas, 64 skills, 5 kits, 19 fix pages; typecheck: 0 errors; npm test: 26/26 passed; next build: 790 static pages generated).
  • Not run: site Playwright smoke suite (npm run smoke) and experience-architecture browser subtests.
Open in Web Open in Cursor 

Co-authored-by: Atlas <AtlasScope@users.noreply.github.com>
@vercel

vercel Bot commented Nov 19, 2026 •

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated
agent-kit Ready Ready Preview Nov 19, 2026 3:35pm UTC

Request Review

@cursor

cursor Bot commented Nov 19, 2026

Copy link
Copy Markdown

Re-review at 217dc2f: ready for Austin's decision

No must-fix items remain.

Counts are now correct:

  • engineer/SKILL.md reads "56 investigations, 12 thinking prompts and 30 skills", and line 17 says "all 56". This matches build_references.mjs output (56 / 12) and the 30 commands in start-here.md.
  • claude-ai/SKILL.md reads "36 specialist skills, 23 recipes", which is correct.
  • The engineering-setup wording now matches its two questions.
  • Frontmatter and structure are untouched in all three skill files.

The version bumps follow the documented process (AGENTS.md on #88):

  • design-kit 1.4.0 to 1.4.1, engineering-kit 1.6.0 to 1.6.1, and experience-architecture 0.4.0 to 0.4.1.
  • Each is updated in both .claude-plugin/plugin.json and .codex-plugin/plugin.json, and in .claude-plugin/marketplace.json.
  • library/library.json badges: 13 engineering-kit and 3 experience-architecture. design-kit badges carry no version.
  • No stale manifest version is left behind. The remaining 1.4.0, 1.6.0 and 0.4.0 strings are upstream skill metadata and a test fixture.
  • .agents/plugins/marketplace.json has no version fields. The experience-architecture package.sh reads its version from plugin.json.

Scope: the diff is limited to plugin READMEs, the three skill text edits, version fields, badges, and library/dist/ (inventory.csv and prompt-library.html). There are no code, script or hook changes.

Generated files: rerunning node library/build.mjs and all three router-reference builders on this branch leaves the tree unchanged, so the committed dist is reproducible.

Checks:

  • Locally: check-public, check-catalog (5 plugins, 64 skills, catalogs agree), library/check (254 entries) and the routing cases all pass.
  • CI is green and ran every step, including the site build and browser smoke tests.

New problems: none found. Plugin README paths are accurate:

  • Run from kits/design-kit and the zip path.
  • kits/design-kit/scripts/impeccable-helpers.json.
  • The full --plugin-dir path for experience-architecture.

The engineering-kit plugin README still has 40+ words outside code blocks.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants