Content
52%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill has a well-defined workflow and provides actionable guidance with concrete commands and output templates, but it is severely over-long and verbose. It repeatedly explains concepts Claude would understand implicitly (intent parsing, anti-patterns, self-checks) and inlines massive formatting templates that should be in separate reference files. The content would benefit enormously from being split into a concise overview SKILL.md with supporting reference files.
Suggestions
Cut content by 50-60%: Remove redundant anti-pattern warnings, self-check reminders, and explanations of things Claude already knows (e.g., how to parse user intent, how to remember context). The 'CONTEXT MEMORY' section and repeated 'CRITICAL' callouts add bulk without value.
Split into multiple files: Move output format templates (stats display, footer formats) into a FORMATS.md, query type search strategies into a QUERIES.md, and setup instructions into a SETUP.md. Reference them from a concise overview.
Add error handling for the research script: What should Claude do if the Python script fails, returns empty results, or times out? This validation gap prevents workflow_clarity from scoring higher.
Remove the marketing/promo banner instructions and API key upsell messaging — these are not skill instructions and waste tokens on every invocation.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose at ~350+ lines. Extensively explains intent parsing, query type classification, and output formatting that Claude could infer from much shorter instructions. Repeated anti-pattern warnings, self-check reminders, and redundant formatting templates bloat the content significantly. The 'CONTEXT MEMORY' section tells Claude to remember things it would naturally remember. | 2 / 5 |
Actionability | Provides concrete bash commands for setup, a specific script invocation pattern, exact output format templates, and a clear quality checklist. The research script call is executable. Minor gap: the script path references a bundle file not provided, and some WebSearch guidance is more directional than executable (e.g., 'choose search queries based on QUERY_TYPE' with example queries but no actual tool invocation syntax). | 4 / 5 |
Workflow Clarity | The multi-step workflow is clearly sequenced: parse intent → setup check → run script → WebSearch → synthesize → display → wait for user → write prompt → follow-up. Steps are numbered and ordered. However, there's no explicit validation/error-handling for the research script failing, and the 'wait for background script' step is vaguely placed. The feedback loop for prompt quality (checklist) is good but validation of research output is implicit. | 4 / 5 |
Progressive Disclosure | Everything is crammed into a single monolithic SKILL.md with no references to supporting documentation. The output format templates, query type classification logic, setup instructions, and prompt writing guidelines could all be split into separate files. No bundle files are provided despite the skill referencing a script at ~/.claude/skills/last30days/scripts/last30days.py. The file is far too long for a single skill document. | 2 / 5 |
Total | 12 / 20 Passed |