Content
75%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, actionable skill body: concrete MCP and CLI usage, a clear three-step workflow with error-recovery guidance, and real one-level-deep bundle references. Weaknesses are minor — some duplication between Overview/Prerequisites/Configuration, placeholder paths instead of a worked example, and references listed rather than linked where they are needed.
Suggestions
Remove the redundant Configuration table (or the Prerequisites bullet) — both state the CONTEXT7_API_KEY requirement, and the Overview paragraph also restates the description.
Provide one fully worked example with real values (e.g. resolving 'react' and querying its docs) so commands and MCP calls are copy-paste ready without placeholders.
Link references at the point of need: in the Step 3 failure path and CLI fallback sections, point to references/troubleshooting.md rather than only listing it under Resources.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient — tool signatures, CLI commands, and a workflow without explaining concepts Claude already knows. Minor trimmable redundancy: the Overview paragraph largely repeats the frontmatter description, and the Configuration table restates the CONTEXT7_API_KEY prerequisite already covered in Prerequisites. Not a 5 for that duplication; well above the padded verbosity of lower anchors. | 4 / 5 |
Actionability | Concrete MCP call signatures with argument formats (`resolve_library_id(query="...", libraryName="...")`), copy-paste CLI commands (`uv run --script {path_to_this_skill}/scripts/context7_cli.py query-docs <library_id> <query>`), an executable bash key-existence check, and a sample output format. It stops short of 5 because `{path_to_this_skill}` and bare `"..."` placeholders mean commands are not literally copy-paste ready and no filled-in worked example is given. | 4 / 5 |
Workflow Clarity | The three-step workflow (check availability → resolve ID → query docs) is clearly sequenced with decision points at each step: ask the user if ambiguous, state no-match with refinement suggestions, and web-fetch the source URL if results are insufficient, plus a retry cap of three attempts. It is not 5 because there is no explicit validation of the final answer against the user's version context, and these error-recovery paths are stated but not structured as validate→fix→retry checkpoints. | 4 / 5 |
Progressive Disclosure | The ~95-line body is an appropriate overview, and both referenced bundle files (`scripts/context7_cli.py`, `references/troubleshooting.md`) exist and are exactly one level deep, clearly listed in a Resources section. Not a 5 because the references are only listed at the bottom rather than signaled at the point of need — e.g. the workflow's failure paths never link to troubleshooting.md, and the tool sections could link the CLI script inline. | 4 / 5 |
Total | 16 / 20 Passed |