Content
81%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.
The body is a well-sequenced, actionable decision workflow: concrete search commands, an explicit vetting checklist, ranking criteria, and a user-decision gate. Its main deductions are small — a few trimmable lines and no fallback guidance for unavailable search tools.
Suggestions
Remove the "Source: salvaged from stale community PR #1232 by redminwang" provenance line and the Related section listing non-bundled skills — neither helps execution and both cost tokens.
Add a brief fallback for environments without gh or web search (e.g., "if gh is unavailable, use web search with the site:github.com query") to make the remote-search step robust.
Consider moving the full example tables into a short examples reference so SKILL.md stays a lean overview if the skill grows.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence — commands are given without library tutorials — but trimmable remnants remain, e.g. "Source: salvaged from stale community PR #1232 by redminwang" and a Related section listing three skills not bundled with this one. Anchor 4 (efficient, minor trimmable instances) fits better than 5. | 4 / 5 |
Actionability | Concrete, mostly copy-paste-ready commands (find/grep/gh search) plus explicit ranking rules, a decision-options table, and worked example tables. Minor gaps: search commands rely on placeholder "keyword|synonym" substitution and there is no fallback when gh or web search tools are unavailable — anchor 4 rather than fully-ready anchor 5. | 4 / 5 |
Workflow Clarity | Six clearly sequenced steps with an explicit validation checkpoint (Step 4's vetting checklist: read frontmatter, look for unexpected commands/network calls, check maintenance) before any adoption, plus a decision gate in Step 6 ("Only create a new skill after the user chooses that path"). The searches are read-only, so no destructive-operation validation is missing. | 5 / 5 |
Progressive Disclosure | No bundle files exist, and the single file is well-sectioned with clear headers and inline examples that are easy to navigate. However, at ~135 body lines it exceeds the under-50-line simple-skill case, and the full example tables (Result Table, User-Facing Summary) could be split out — good structure with minor organization gaps, matching anchor 4. | 4 / 5 |
Total | 17 / 20 Passed |