Content
85%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 a well-structured, evidence-gated workflow with concrete connector commands and clean progressive disclosure of detail into references. The main opportunities are trimming the repeated Proxy/NOT_SCORED guardrail and surfacing a worked success example inline rather than only the NEEDS_INPUT case.
Suggestions
State the Proxy candidate / NOT_SCORED rule once authoritatively and reference it from later steps instead of restating it in Data Sources, each Instruction step, and the Example.
Add a brief inline worked success snippet (one scored candidate with a go call) so the common happy path is visible without opening references/templates.md.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and assumes Claude's competence — no padding about what trends or platforms are — but the Proxy candidate / NOT_SCORED gate is restated across Data Sources, Instructions, and the example, which is minor redundancy that could be tightened. | 4 / 5 |
Actionability | Provides copy-paste connector commands (tavily.py, pageviews.py), the HN Algolia API URL, exact save paths, and explicit field lists, but the inline Example demonstrates only the NEEDS_INPUT path with the full worked success case living one level deep in references. | 4 / 5 |
Workflow Clarity | A 7-step numbered sequence with explicit validation gates (current-evidence gate), a NEEDS_INPUT feedback loop for error recovery, a 'Done when' checklist, and separate-authorization gating before ledger writes. | 5 / 5 |
Progressive Disclosure | SKILL.md is a lean overview pointing one level deep to references/templates.md (the long scaffolds) and references/trend-scout-recipe.md (the keyless recipe), with clearly signaled links and easy navigation. | 5 / 5 |
Total | 18 / 20 Passed |