Content
71%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.
Highly actionable and clearly sequenced: a user or agent could execute the analysis immediately from the exact queries, URLs, and output template provided. The weaknesses are redundancy (duplicated intro and agent-integration workflow) and a monolithic single-file structure where the report template and keyword lists belong in reference files.
Suggestions
Move the ~55-line markdown report template (Phase 7) into a references/report-template.md file and replace it with a one-line pointer, cutting a large chunk of inline tokens.
Delete or merge the "AI Agent Integration" section into the Process overview — its 6-step workflow duplicates the seven phases already documented, and the example prompt repeats the Quick Start example.
Remove the opening paragraph that restates the frontmatter description nearly verbatim, and add per-phase guidance for handling empty or blocked search results (e.g., query variations, retry cadence) to strengthen workflow checkpoints.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The query lists are lean, but the opening paragraph restates the frontmatter description almost verbatim, the ~55-line report template is inlined wholesale, and the "AI Agent Integration" section re-describes the Process phases already covered above. | 3 / 5 |
Actionability | Guidance is copy-paste ready throughout: exact search operators ("site:[domain] filetype:pdf"), concrete WebFetch URLs ("https://www.similarweb.com/website/[domain]/"), per-category keyword patterns, and a worked example with real domains (pump.co, vantage.sh). | 5 / 5 |
Workflow Clarity | Seven phases are clearly sequenced with per-phase extraction targets, and a fallback is noted for SimilarWeb rate-limiting; however, no explicit per-phase validation checkpoints exist (e.g., what to do when a search returns nothing), keeping it below a 5. The skill is read-only, so the destructive/batch validation cap does not apply. | 4 / 5 |
Progressive Disclosure | The single file is well-sectioned, but at 230+ lines with the full report template and keyword taxonomies inlined — content that clearly belongs in separate reference files — and no bundle files or references exist, organization could be significantly better. | 3 / 5 |
Total | 15 / 20 Passed |