Content
86%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 highly actionable, well-structured, and uses progressive disclosure cleanly with real one-level-deep references matched to task triggers. It is only marginally held back by a little orienting prose and a few workflows lacking explicit validate→retry checklists on destructive/batch paths.
Suggestions
Tighten the introductory orientation (the trunk-first diagram and up/down semantics) to lift conciseness toward a 5.
Convert the merge and sync-divergence guidance into explicit validate→decide→retry sequences to give those batch/destructive paths the same checkpoint rigor as the exit-3 recovery.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and mostly assumes Claude's competence — no padded explanations of git, PRs, or libraries — with a few sentences of orientation prose (the trunk-first diagram and up/down semantics) that could be trimmed slightly, keeping it just below a 5. | 4 / 5 |
Actionability | Copy-paste-ready commands throughout (Setup, Core loop, Branch placement, Staying in sync, Merging) plus a concrete always/never-run table and an exit-code recovery map covering the common cases fully. | 5 / 5 |
Workflow Clarity | Core loop and branch-placement flows are explicitly sequenced with verification ("gh stack view --json" to confirm) and an exit-3 recovery loop, but a couple of batch/destructive paths (merge, sync divergence) rely on prose rather than an explicit validate→fix→retry checklist, so it stops just short of a 5. | 4 / 5 |
Progressive Disclosure | SKILL.md is a tight overview that keeps each topic inline-only as needed and points to three clearly signaled, one-level-deep references (stack-design.md, commands.md, troubleshooting.md) each with a trigger-matched navigation cue; the referenced files all exist under references/. | 5 / 5 |
Total | 18 / 20 Passed |