Content
93%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 tight, highly actionable skill body that leads with concrete call signatures and organizes guidance by triggering scenario. The only soft spot is a couple of single-call sections that lack an explicit verify/recover step, though the skill is non-destructive so the workflow cap does not apply.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence — no padding, no explaining what an ADR or centrality is — and every line (the four get_why modes, the source list, the CLI commands) earns its place. | 5 / 5 |
Actionability | It gives copy-paste-ready, concrete calls for every common case — get_why(query=...), targets=[...], get_why(), repowise decision health/confirm/add, repowise why — covering the main cases with executable syntax. | 5 / 5 |
Workflow Clarity | The 'When about to make an architectural change' section is a clear 3-step sequence with an explicit checkpoint ('present them to the user before proceeding'), but other sections rely on a single call without an explicit verify/feedback step, leaving minor checkpoint gaps relative to the 5 anchor. | 4 / 5 |
Progressive Disclosure | No bundle files exist and none are needed; the body is well under 50 lines and cleanly organized into purpose-labeled sections ('When the user asks…', 'When about to make…', 'When called with no specific query'), so the simple-skill exception grants a 5. | 5 / 5 |
Total | 19 / 20 Passed |