diff --git a/.github/workflows/checks.yml b/.github/workflows/checks.yml index 66cab9c..a05961d 100644 --- a/.github/workflows/checks.yml +++ b/.github/workflows/checks.yml @@ -10,8 +10,10 @@ jobs: steps: - name: Checkout the repository uses: actions/checkout@v7 - - name: Check editorconfig + - name: Install editorconfig-checker uses: editorconfig-checker/action-editorconfig-checker@main + - name: Check editorconfig + run: editorconfig-checker package-json: runs-on: ubuntu-latest name: package.json diff --git a/skills/codex/SKILL.md b/skills/codex/SKILL.md index 9708a52..41f2040 100644 --- a/skills/codex/SKILL.md +++ b/skills/codex/SKILL.md @@ -132,16 +132,16 @@ Codex is powered by OpenAI models with their own knowledge cutoffs and limitatio When Codex is wrong: -1. State your disagreement clearly to the user. -2. Provide evidence: your own knowledge, web search, docs. -3. Optionally resume the Codex session to discuss the disagreement. Be transparent about the sender: disclose your own actual model name, so Codex knows the follow-up comes from the AI agent coordinating the session rather than from the human user: +1. State your disagreement clearly to the user. +2. Provide evidence: your own knowledge, web search, docs. +3. Optionally resume the Codex session to discuss the disagreement. Be transparent about the sender: disclose your own actual model name, so Codex knows the follow-up comes from the AI agent coordinating the session rather than from the human user: - ```bash - echo "Follow-up from , the AI agent coordinating this session. I disagree with [X] because [evidence]. What's your take on this?" | codex exec --skip-git-repo-check resume --last 2>/dev/null - ``` + ```bash + echo "Follow-up from , the AI agent coordinating this session. I disagree with [X] because [evidence]. What's your take on this?" | codex exec --skip-git-repo-check resume --last 2>/dev/null + ``` -4. Frame disagreements as discussions, not corrections — either AI could be wrong. -5. Let the user decide how to proceed if there's genuine ambiguity. +4. Frame disagreements as discussions, not corrections — either AI could be wrong. +5. Let the user decide how to proceed if there's genuine ambiguity. ## Error Handling diff --git a/skills/project/SKILL.md b/skills/project/SKILL.md index abd2fbc..b169e91 100644 --- a/skills/project/SKILL.md +++ b/skills/project/SKILL.md @@ -32,15 +32,15 @@ Generate a project base from reference repositories: pick the closest references ## Process -1. **Clarify the project profile if the user hasn't specified it.** You need four answers before generating: - - language: TypeScript or JavaScript; - - runtime: Node.js, frontend (browser), or both; - - structure: single package or pnpm monorepo; - - kind: published library or GitHub action (affects publishing/release setup). -2. **Pick 1–2 references** from the table below that best match the profile. -3. **Fetch the actual files** from the reference repos (e.g. `https://raw.githubusercontent.com///HEAD/`, or `gh api`) and use them as the source. References evolve; the fetched file wins over anything written in this skill. -4. **Adapt** the copied files: project/package names, org and repo URLs, description, author, commitlint scopes, Node.js/pnpm versions (keep workflows in sync with `.tool-versions`). Drop files irrelevant to the profile (e.g. `pnpm-workspace.yaml` for a single package, `.size-limit.json` when bundle size doesn't matter). -5. **Finish up:** `pnpm install`, `pnpm updateGitHooks`, then verify `pnpm test` passes. +1. **Clarify the project profile if the user hasn't specified it.** You need four answers before generating: + - language: TypeScript or JavaScript; + - runtime: Node.js, frontend (browser), or both; + - structure: single package or pnpm monorepo; + - kind: published library or GitHub action (affects publishing/release setup). +2. **Pick 1–2 references** from the table below that best match the profile. +3. **Fetch the actual files** from the reference repos (e.g. `https://raw.githubusercontent.com///HEAD/`, or `gh api`) and use them as the source. References evolve; the fetched file wins over anything written in this skill. +4. **Adapt** the copied files: project/package names, org and repo URLs, description, author, commitlint scopes, Node.js/pnpm versions (keep workflows in sync with `.tool-versions`). Drop files irrelevant to the profile (e.g. `pnpm-workspace.yaml` for a single package, `.size-limit.json` when bundle size doesn't matter). +5. **Finish up:** `pnpm install`, `pnpm updateGitHooks`, then verify `pnpm test` passes. ## References @@ -96,16 +96,16 @@ All Node.js jobs: `pnpm/action-setup` → `actions/setup-node` with `cache: 'pnp Files to look for in references and generate, grouped by concern: ``` -Toolchain: .tool-versions .editorconfig .gitignore -Packages: package.json pnpm-workspace.yaml packages/*/package.json -TypeScript: tsconfig.json tsconfig.build.json env.d.ts -Lint: oxlint.config.* .npmpackagejsonlintrc.json -Commits: .commitlintrc.* .czrc .simple-git-hooks.json .nano-staged.json -Build/test: vite.config.* .size-limit.json -Release: .clean-publish .simple-release.* -GitHub: .github/workflows/{checks,tests,commit,release}.yml .github/renovate.json - .github/ISSUE_TEMPLATE/{config,bug-report,feature-request,question}.yml .github/FUNDING.yml -Docs: README.md (badges: npm, node, dependencies, build) LICENSE (MIT) +Toolchain: .tool-versions .editorconfig .gitignore +Packages: package.json pnpm-workspace.yaml packages/*/package.json +TypeScript: tsconfig.json tsconfig.build.json env.d.ts +Lint: oxlint.config.* .npmpackagejsonlintrc.json +Commits: .commitlintrc.* .czrc .simple-git-hooks.json .nano-staged.json +Build/test: vite.config.* .size-limit.json +Release: .clean-publish .simple-release.* +GitHub: .github/workflows/{checks,tests,commit,release}.yml .github/renovate.json + .github/ISSUE_TEMPLATE/{config,bug-report,feature-request,question}.yml .github/FUNDING.yml +Docs: README.md (badges: npm, node, dependencies, build) LICENSE (MIT) ``` `CHANGELOG.md` is generated by simple-release — don't create it by hand.