Content
92%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 lean, highly actionable diagnostic skill body with concrete commands, explicit validation/feedback workflows, and well-organized one-level-deep reference files. Slight conciseness dip comes from thematic repetition of the "re-read from scratch" directive.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense and operational throughout — it assumes Claude's competence (no explaining what git bisect or a debugger is) and every section is actionable rules rather than padded explanation. Minor repetition of the "re-read the execution path from scratch" theme across Hard Rules, Confirm or Discard, and rationalization smells keeps it just below the lean 5 anchor. | 4 / 5 |
Actionability | Fully executable guidance with concrete commands ("git status --short --branch -uall", "git diff <last-good>..HEAD -- <suspect path>", "git bisect good/bad", "sample <process>", "sw_vers", "grep -rn"), specific file:line examples, and copy-paste output/handoff templates covering the common cases. | 5 / 5 |
Workflow Clarity | Multi-step processes are explicitly sequenced with validation checkpoints (Runtime Evidence Ladder 1–5, Native App Freeze evidence 1–4, Bisect Mode, regression guard red-green run), plus feedback loops (Confirm or Discard, hard stops, 3-hypothesis handoff) — no destructive/batch cap applies since validation is pervasive. | 5 / 5 |
Progressive Disclosure | Clear overview body with five well-signaled one-level-deep references ("See references/durable-context.md", "Load references/failure-patterns.md", etc.), all of which exist and contain no nested references — content appropriately split with easy navigation. | 5 / 5 |
Total | 19 / 20 Passed |