Content
56%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 highly actionable with concrete configs, commands, examples, and thoughtful gotchas/safety guidance, and the workflow is clear for a simple single-task skill. Its main problems are pervasive duplication of parameters, endpoints, and field lists across sections, and reference-style content inlined in an overlong single file.
Suggestions
Consolidate the parameter/endpoint listings (frontmatter, Input Formats, Tool call, CLI flags, MCP Reference) into a single canonical reference section and remove the repeats.
Trim or remove 'Why This Exists' and 'Algorithm/Methodology' — the agent operating the skill does not need the server's ingestion pipeline or marketing rationale.
Move the full Example Output, Output Structure field list, Pricing, and Citations into a references/ file (e.g. references/output-format.md) linked one level deep from SKILL.md.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is noticeably padded: the same three parameters are listed four times (frontmatter inputs, Input Formats table, Tool call block, CLI flags table), endpoints appear in four places, '25+ fields' is repeated ~6 times, and sections like 'Why This Exists' and 'Algorithm/Methodology' (server internals) add little operational value. | 2 / 5 |
Actionability | Provides copy-paste-ready MCP client config JSON, an npx alternative, a concrete tool-call spec with params, CLI commands with a flag table, and a demo with expected output — minor gaps only, e.g. 'python clawbio.py run' presumes a runner whose availability is never established. | 4 / 5 |
Workflow Clarity | A clear 5-step workflow with freedom-level guidance, explicit failure handling ('If the MCP call fails or returns no results, say so. Never invent titles, DOIs, or findings'), rate-limit handling, and a separate demo verification path — though no validation checkpoint is embedded in the workflow itself. | 4 / 5 |
Progressive Disclosure | A single ~270-line file with clear section headers but no bundle files; bulk reference-style content (full Example Output, Output Structure field list, Pricing, Citations) is inlined rather than split into one-level-deep reference files. | 3 / 5 |
Total | 13 / 20 Passed |