Content
63%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 thorough, actionable instruction skill with clear sequenced workflows and concrete MCP tool calls, weakened by verbosity in the config section and by relying on five reference files that are not present in the bundle.
Suggestions
Tighten the Run posture section: move the per-field prose for lower-frequency config options (write_scopes, output_destinations, mcp_gateway_server_ids, repositories) into a reference file, keeping only the common fields and the discriminator guidance inline.
Ship the five referenced files (references/scout-anatomy.md, scout-patterns.md, report-contract.md, dedupe-and-memory.md, lifecycle-and-testing.md) in the bundle, since critical content like the frontmatter schema and report-contract bar is delegated to them and the links are currently dead.
In the Test loop, make the validation gates explicit and ordered (dogfood queries pass → create → run-now → inspect) as a numbered checklist with a hard 'do not spend a run before dogfooding' checkpoint, rather than advisory prose.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and on-topic (scout-specific, not general knowledge Claude already has), but the Run posture section runs very long with multi-paragraph field descriptions and repeated cautions that could be tightened or pushed to references; it fits 'mostly efficient but could be tightened' rather than the 4 anchor's 'minor instances'. | 3 / 5 |
Actionability | Gives concrete, executable guidance — named MCP tools with JSON params (e.g. 'posthog:scout-run-now {"id": <config_id>}', 'posthog:scout-create'), exact field ranges (run_interval_minutes 30–43200) and scope names — with only minor gaps in fully-specified call signatures, fitting the 4 anchor. | 4 / 5 |
Workflow Clarity | Multi-step flows are clearly sequenced with checkpoints — 'The job before the writing' (4 numbered steps), decision tables, and a dogfood → run-once → inspect test loop with feedback (refine → re-query → repeat); not a 5 because some validation is advisory rather than a hard gate and the workflow is spread across sections. | 4 / 5 |
Progressive Disclosure | The body is well structured with five references that are one-level-deep and clearly signaled via markdown links plus one-line content descriptions, but none of the referenced files (references/scout-anatomy.md, scout-patterns.md, report-contract.md, dedupe-and-memory.md, lifecycle-and-testing.md) exist in the bundle, so the disclosure structure cannot be verified and critical content (frontmatter schema, report contract) is delegated to missing files. | 3 / 5 |
Total | 14 / 20 Passed |