Content
80%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: executable BAD/GOOD anti-pattern examples, a clean one-level-deep reference table verified against the actual bundle, and tight scope boundaries with explicit sibling-skill routing. Its weaknesses are the absence of an explicitly sequenced usage workflow (selection → reference lookup → implementation → type-check is only implicit) and mild redundancy between the 'When to Apply' and 'Use When' sections.
Suggestions
Replace or merge the overlapping 'When to Apply' and 'Use When' sections into one list pairing each trigger question with its reference file, which would simultaneously fix the redundancy and make the selection workflow explicit.
Add a short 3–4 step usage workflow ending in the existing `npx tsc --noEmit` validation step (pick pattern → read its reference file → implement → type-check) so the sequence and checkpoint are explicit rather than implied.
Trim the Mindset paragraph to its actionable core (the runtime-failure-to-compile-error test and the unconstrained-generic rule), cutting roughly a third of its prose.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with opinionated, non-obvious guidance ('a builder that still lets `.build()` run with missing fields is a fluent API, not a type-safe one') and never explains concepts Claude already knows, but the 10-bullet 'When to Apply' list and the 4-question 'Use When' list overlap noticeably, and the Mindset prose could be tightened — minor over-explanation fitting anchor 4 rather than the every-token-earns-its-place anchor 5. | 4 / 5 |
Actionability | Three complete, compilable BAD/GOOD TypeScript pairs (builder with `build(this: UserBuilder<true>)`, constrained `createInstance` using `ConstructorParameters<T>`/`InstanceType<T>`, typed emitter with `Parameters<T[K]>`) plus the runnable command `npx tsc --noEmit` — copy-paste ready code covering the most common misuse cases, matching anchor 5. | 5 / 5 |
Workflow Clarity | No multi-step sequence is explicitly stated — the implied flow (match the scenario in 'When to Apply', consult the relevant reference file, implement, verify) is implicit, and while the type-check validation checkpoint is provided, routing guidance and steps are never laid out as a sequence, matching anchor 3 ('sequence present but implicit') rather than anchor 4's clearly listed steps. | 3 / 5 |
Progressive Disclosure | A clear overview with 12 one-level-deep references presented in a table with per-file 'Covers' descriptions, all of which exist on disk; pattern details correctly live in the reference files rather than inline, and the inline anti-patterns are appropriate overview-level material — matching anchor 5's well-signaled, easy-to-navigate structure. | 5 / 5 |
Total | 17 / 20 Passed |