Content
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured, audit-first instruction skill with a clear sequenced workflow and concrete output template. Its main weakness is mild redundancy where the live-state taxonomy and guardrail cautions are restated across sections.
Suggestions
Define the five live-state categories once (e.g., in Workflow step 2) and reference them from Guardrails instead of repeating the full list.
Collapse overlapping cautions: 'do not claim a tool is live just because a skill or config references it' (Guardrails) and 'do not treat present in config as working' (Pitfalls) say the same thing — keep one.
Optionally add one or two concrete enumeration commands (e.g., `gh workflow list`, inspecting `.mcp.json` / hook config) to push actionability from strong-guidance to fully executable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean and bullet-driven, but the five live-state categories (configured / authenticated / recently verified / stale or broken / missing) are repeated in both Guardrails and Workflow step 2, and Pitfalls restates Guardrails. | 2 / 3 |
Actionability | For an instruction-only orchestration skill it gives concrete surfaces to inventory, explicit classification categories, a proof-path source list, and a copy-paste ready output format template, delegating enumeration to named sibling skills. | 3 / 3 |
Workflow Clarity | A clearly sequenced four-step workflow (Inventory -> Classify -> Trace proof -> Recommend) with an explicit ambiguity checkpoint ('If the current state is ambiguous, say so directly') and a dedicated Verification section. | 3 / 3 |
Progressive Disclosure | Well-organized into focused sections with one-level-deep references to sibling ECC skills (workspace-surface-audit, github-ops, etc.); no nested or monolithic references, and no bundle files to misroute. | 3 / 3 |
Total | 11 / 12 Passed |