← ClaudeAtlas

cupertino-handbook-checklisted

Use when the user wants to compare new or changed work against an existing domain handbook at .cupertino/DOMAIN-handbook.md (one file per domain) to find divergence. Trigger on 'does this comply with our handbook', 'audit these files against the code handbook', 'check this against the code handbook'. Requires a handbook previously produced by cupertino-handbook-draft. Not for repo-wide conventions written in CLAUDE.md or .claude/house-rules.md — use self-assess:self-assess-lint-audit for those. Read-only — never modifies target files. Zero findings is a valid, expected outcome, not something to work around.
Anselmoo/werkstoff · ★ 1 · Code & Development · score 67
Install: claude install-skill Anselmoo/werkstoff
Check files against an existing handbook, one rule at a time, verifying every finding independently. Never touch the target files. ## Steps 1. **Parse the domain**, confirm `.cupertino/<domain>-handbook.md` exists (if not, point the user to `cupertino-handbook-draft`). 2. **Determine target files**: use the files the user named, or if they asked for a check but didn't say what changed, **ask whether to run a full-project scan** rather than silently picking a scope for them. 3. **Mark the check as active** so the write/no-mutation guarantees below are enforced at the tool-call layer, not just by convention: ```bash python3 "${CLAUDE_PLUGIN_ROOT}/scripts/state.py" set handbook-check-active ``` 4. **Read the handbook** and extract its rule list (one `{dimension, rule}` per handbook entry). 5. **Run the check** via the Workflow tool: ``` Workflow({ scriptPath: "${CLAUDE_PLUGIN_ROOT}/workflows/handbook-check.js", args: { rules: [...], targetFiles: [...] } }) ``` This dispatches `handbook-drift-auditor` once per rule (never bundling rules), then independently re-verifies every individual finding at its exact `file:line` before it counts — a PreToolUse hook backstops both the one-rule-per-dispatch and one-location-per-verify constraints regardless of what the workflow script does. Findings that don't survive re-verification are dropped, not reported. 6. **Clear the active flag** when done — and immediately, before stopping, if step 4 or step 5 fails or is inte