Content
65%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 overview skill with excellent progressive disclosure and concrete commands/snippets, weakened by duplicated trigger sections ('When to Apply' vs 'Use When') and by guidance that stays at the heuristic level for its multi-step tasks — particularly the slow-type-checking diagnosis it names as a core use case. Tightening the redundancy and giving the diagnosis path an explicit, checkpointed sequence would lift it substantially.
Suggestions
Merge 'When to Apply' and 'Use When' into a single section — they currently list the same four triggers in slightly different words — and drop the 'In Scope' list that restates the frontmatter description.
Replace the 'profile before guessing' heuristic with an explicit sequenced diagnostic workflow: run 'tsc --noEmit --extendedDiagnostics', check the reported metrics (files, instantiation count, memory), then branch to the matching fix (e.g., project references, isolating deep conditional-type recursion), re-running diagnostics to confirm the gain.
Add one complete recommended tsconfig.json for the most common case (e.g., a Node/ESM project with the full strictness flag set) so the bootstrap use case is copy-paste ready rather than assembled from the anti-pattern GOOD snippets.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body assumes Claude's competence (no 'what is TypeScript' padding), but 'When to Apply' and 'Use When' are near-duplicate lists of the same four triggers, and 'In Scope' restates the frontmatter description — roughly a dozen redundant lines. That is more than the 'minor instances' of the score-4 anchor; it matches the score-3 anchor ('mostly efficient but... could be tightened'), while remaining well above the verbose score-2 anchor. | 3 / 5 |
Actionability | Concrete, executable material throughout: three runnable commands ('npx tsc --noEmit --extendedDiagnostics', 'npx tsc --init') and complete BAD/GOOD tsconfig.json snippets. Minor gaps keep it below the score-5 anchor: no full recommended tsconfig for the common Node/ESM bootstrap case and no worked example of acting on extendedDiagnostics output, so not every common case is copy-paste covered. | 4 / 5 |
Workflow Clarity | The performance path has a sensible implicit sequence ('Run tsc --noEmit --extendedDiagnostics first, then act on what it reports') but no explicit ordered steps or checkpoints for interpreting diagnostics, and the bootstrap/module-resolution tasks have no sequence at all — they just point at references. This matches the score-3 anchor ('sequence present but checkpoints missing or implicit') rather than score 4, which requires most checkpoints to be explicit; no destructive/batch cap applies. | 3 / 5 |
Progressive Disclosure | The body is a genuine overview: mindset, scope, commands, and anti-patterns inline, with a clean references table mapping each of the four topics to a real one-level-deep file (all four referenced files exist in references/) plus official docs. This matches the score-5 anchor ('clear overview with well-signaled one-level-deep references; content appropriately split; easy navigation') and exceeds the score-4 anchor, which allows organization gaps. | 5 / 5 |
Total | 15 / 20 Passed |