Content
86%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 skill body: a clear method-fallback workflow, untrusted-content handling, and proper progressive disclosure across four real reference files. The two dimensions at 4 reflect small gaps — no example Mermaid/classDef output and no validation that the generated diagram renders.
Suggestions
Add a short example Mermaid snippet (with classDef styling) in the Formatting Guidelines so the color rules translate directly to executable output.
Add a final validation step in Step 4 — e.g., check the generated diagram for syntax errors or unexpected node labels before returning it.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body assumes Claude's competence throughout — it never explains what dbt, Mermaid, or MCP tools are — and every line is instruction or decision guidance (e.g., "Check file size first — if too large (>10MB), skip to next method"). Phrases like "Provides production lineage from dbt Cloud" disambiguate method choice rather than pad, matching the lean 5 anchor. | 5 / 5 |
Actionability | Guidance is largely executable: a four-method hierarchy with links, the "graph LR" directive, and an exact color-to-node-type mapping. However, no example Mermaid snippet or classDef/styling syntax is given for actually applying those colors, leaving a minor gap between the concrete instructions and copy-paste-ready output. | 4 / 5 |
Workflow Clarity | Steps 1–4 are clearly sequenced, the fallback chain is explicit, and there is one conditional checkpoint (the >10MB manifest size check). A validation step for the generated output — e.g., confirming the Mermaid renders without syntax errors — is missing, which keeps it at 4 rather than 5. | 4 / 5 |
Progressive Disclosure | The body is a lean overview with four clearly signaled, one-level-deep references (using-get-lineage-dev.md, using-get-lineage.md, using-manifest-json.md, parsing-code-directly.md), all verified to exist with no nested references. ~310 lines of method detail are appropriately split out of SKILL.md, exactly matching the top anchor. | 5 / 5 |
Total | 18 / 20 Passed |