Content
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-crafted skill with excellent actionability and workflow clarity — the five-phase process is clearly sequenced, tool usage is specific, and the examples effectively demonstrate the expected behavior. Its main weakness is verbosity: Phase 1 explains question-understanding techniques Claude already possesses, the search tips partially duplicate phase instructions, and four full examples with identical structure add tokens without proportional insight. The skill would benefit from trimming redundant guidance and potentially splitting examples into a separate reference file.
Suggestions
Trim Phase 1 significantly — Claude already knows how to extract core questions and infer user context. A single sentence like 'Extract the core question, domain, and key terms before searching' suffices.
Consolidate the 'Search Strategy Tips' into Phase 2 directly, removing duplicated advice about using concrete terms and broadening queries.
Reduce examples from 4 to 2 (they follow an identical pattern), or move them to a separate EXAMPLES.md file to keep the main skill leaner.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is well-written but verbose for what it conveys. Phase 1's bullet points about understanding the question explain things Claude already knows how to do (extracting core questions, identifying domains, inferring user roles). The search strategy tips and troubleshooting sections repeat guidance already covered in the phases. The four detailed examples, while useful, could be more compact — each follows the same pattern and the repetition adds tokens without proportional value. | 2 / 3 |
Actionability | The skill provides highly concrete, executable guidance: specific MCP tool names (search_content, read_content, read_excerpt, list_content), exact search query examples ('pricing AI product outcomes', 'stalled growth logo retention'), clear output format with source citation templates, and four worked examples showing the full search-to-answer pipeline. Every phase has specific, actionable instructions. | 3 / 3 |
Workflow Clarity | The five-phase workflow is clearly sequenced with explicit decision points: how many searches to run, when to use full read vs excerpt, how many pieces to read (2-4 cap), and a structured output format. The troubleshooting section provides error recovery paths for common failure modes (no results, tangential results, broad questions, conflicting advice). The gotchas section adds validation constraints (don't cite unread content, don't read more than 4 pieces). | 3 / 3 |
Progressive Disclosure | The content is a single monolithic file with no references to supporting documents, which is acceptable given no bundle files exist. However, at ~200+ lines, the examples and troubleshooting sections could reasonably be split into separate files. The content is well-organized with clear headers, but the length pushes against what should be inline in a SKILL.md overview. | 2 / 3 |
Total | 10 / 12 Passed |