Content
71%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured orchestrator skill with clear precedence rules, concrete routing, and a sensible findings taxonomy, though slightly verbose in the trigger list. The main defect is that the routing workflow points at six agent files that do not exist in the bundle, leaving the core instructions unresolvable as shipped.
Suggestions
Ship the six referenced agents/*.agent.md files (or remove the agent-mode routing and run all checks inline) so the Step 1 routing targets actually resolve.
Add a post-correction verification step in Step 4 (e.g., re-run the matching convention check on the edited file and confirm no new violations before reporting done).
Trim the 11-item 'Typical triggers' list to the handful that are not already covered by the description's USE FOR clause.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body stays lean and assumes Claude's competence (no concept explanations), but the 11-item 'Typical triggers' list largely duplicates the description's USE FOR list and could be trimmed. Not 5: that list and a few padded phrases are tokens that do not earn their place. | 4 / 5 |
Actionability | Concrete routing by file extension ('TypeScript (.ts, non-component .tsx) -> agents/typescript.agent.md'), an explicit Required/Recommended/Advisory findings taxonomy with example items, and concrete edit constraints ('Don't rewrite entire files unless under 50 lines'). Not 5: no executable examples and the routing delegates to agent files whose content is not present. | 4 / 5 |
Workflow Clarity | Steps 1-4 are clearly sequenced with input-validation gates ('If language cannot be determined, ask before proceeding') and a user-confirmation checkpoint before file mutation in Step 4. Not 5: there is no post-correction verification loop (e.g., re-check that the fix conforms) after applying edits. | 4 / 5 |
Progressive Disclosure | The four references/*.conventions.md and three assets/*.follow-up.md files all exist and are clearly signaled via the agent table and bullet lists, but the six agents/*.agent.md files that the entire routing workflow depends on are absent from the bundle. Not 4: broken references to the primary navigation targets are more than 'minor organization gaps'. | 3 / 5 |
Total | 15 / 20 Passed |