Content
56%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 actionable and well-sequenced with concrete commands, JSON shapes, and validation checkpoints, but it is verbose and its progressive-disclosure references point to protocol files that are absent from the bundle.
Suggestions
Tighten the repeated "Do not..." prohibition lists and governance prose; consolidate duplicated rules to reduce the ~280-line footprint.
Fix the Protocol Map by either adding the referenced protocols/*.md files to the bundle or repointing the links to existing references/ files.
Move the inlined runtime/stage detail into the referenced protocol files so SKILL.md stays a lean overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~280-line body is noticeably verbose: dense governance prose, repeated "Do not..." prohibition lists, and restated rules that could be tightened, even though the embedded commands and JSON carry real content. | 2 / 5 |
Actionability | Provides concrete executable canonical-entry commands (PowerShell and Bash) and a complete host-decision JSON shape with copy-paste-ready structure; only minor gaps from placeholders like <skill_root> and <current user task, verbatim>. | 4 / 5 |
Workflow Clarity | A clear fixed state machine (skeleton_check through phase_cleanup) is sequenced with explicit validation (proof artifacts, verify-before-completion) and a feedback loop (one targeted repair then rerun); minor gaps come from checkpoints being buried in dense prose. | 4 / 5 |
Progressive Disclosure | A well-signaled "Protocol Map" points one level deep to protocols/runtime.md, protocols/think.md, etc., but no protocols/ directory exists in the bundle, so the references are broken; additionally, much detail that belongs in those files is inlined here. | 3 / 5 |
Total | 13 / 20 Passed |