Content
67%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 delivers executable, well-organized search guidance with a clear primary source and a simple workflow appropriate to a single-purpose skill. Its weakest point is token efficiency: generic best-practice advice and a stale-prone skill catalog pad the file without adding capability.
Suggestions
Cut the "Best Practices" and "Installation Tips" sections (or compress to one line each) — guidance like "Read documentation" and "Verify network connectivity" is generic knowledge Claude already has.
Move the "Common Skill Categories" catalog into a references/categories.md file (or drop it) — it is a stale-prone listing that inflates SKILL.md without aiding the search task.
Make the GitHub source actionable: replace "Search: `openclaw skill` or `agent-skill`" with a concrete search URL or command, e.g. `gh search repos "openclaw skill" --filename SKILL.md`.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Command examples are compact, but the body carries sections Claude does not need: "Best Practices" ("Read documentation", "Start simple", "Test thoroughly"), "Installation Tips" ("Check skill requirements", "Verify network connectivity"), and a ~40-line "Common Skill Categories" catalog of skill names that will go stale. It is above 2 because much of the body is efficient executable content, but below 4 due to these padded, low-value sections. | 3 / 5 |
Actionability | Concrete, executable commands dominate: `npx clawhub search "keyword"`, `npx clawhub browse`, `--sort installs`, `--sort stars`, plus direct source URLs. Minor gaps keep it from 5: the GitHub guidance ("Search: `openclaw skill` or `agent-skill`") lacks an executable search command or URL, and troubleshooting steps like "Check skill requirements" are directions rather than commands. | 4 / 5 |
Workflow Clarity | This is a simple single-purpose skill whose primary action (clawhub search, marked "Primary" among sources) is unambiguous, and the numbered Installation Tips sequence (check requirements → read SKILL.md → test in isolation) provides a clear order with a validation checkpoint. It is not 5 because the rate-limit recovery path, while ordered, has no explicit verify-success loop, and the choice among the five sources is only implicitly prioritized. | 4 / 5 |
Progressive Disclosure | The single-file body is well organized with clear section headers and no nested or broken references (no references/, scripts/, or assets/ directories exist). It is not 5 because the file runs ~135 lines and the Common Skill Categories catalog and Search Strategies repetition are content that could be split into a reference file, leaving SKILL.md a leaner overview. | 4 / 5 |
Total | 15 / 20 Passed |