Content
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The content is highly actionable and its workflow is clearly sequenced with strong validation checkpoints for destructive and batch operations. Its main weaknesses are verbosity (duplicated blocks and full display templates that could be condensed) and a monolithic single-file structure with no progressive disclosure into bundle files.
Suggestions
Consolidate the duplicated collision-check logic: --roadmap and the Archive Operation both spell out the same read-then-skip/STOP pattern — extract it into one shared section referenced by both, instead of restating it twice.
Move the large reproduced display/output templates and the Completion Detection Algorithm into reference files under references/ (e.g. DISPLAY_FORMATS.md, COMPLETION.md) and link to them one level deep, reducing the main SKILL.md length and adding genuine progressive disclosure.
Tighten the repeated STOP/error message blocks (Plan not found, not completed, no completed plans) into a compact table or a single enumerated list rather than verbatim fenced blocks per mode.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is accurate and free of concept-explanation padding, but it is verbose — near-duplicate collision-check blocks (in --roadmap and Archive Operation), full reproduced display templates, and repeated STOP/error message blocks that could be tightened; it is "mostly efficient but includes some unnecessary explanation or could be tightened." It is not the level above because true level-3 lean content would consolidate these duplicated blocks rather than restating them per mode. | 2 / 3 |
Actionability | Provides fully executable guidance: concrete commands (mv, mkdir -p, git log -1 --format=%ai), exact Glob patterns, precise frontmatter and comment formats to insert, and copy-paste-ready message templates — matching the "fully executable, copy-paste ready" anchor. Not level 2 because there is no pseudocode or missing key detail. | 3 / 3 |
Workflow Clarity | Clear sequenced steps with explicit validation checkpoints and feedback loops for risky/batch operations: collision checks before moving, "Do NOT edit paths.roadmap unless the snapshot write succeeded", ultra-bundle validation before move, confirmation prompts for --all/--roadmap, and batch skip-vs-single-stop error handling. Not level 2 because validation and error-recovery checkpoints are explicit, not implicit. | 3 / 3 |
Progressive Disclosure | No bundle files exist (references/scripts/assets are empty), so the skill is a monolithic ~340-line single-file SKILL.md with no external references to split out detail; sections are organized but self-contained content like the completion-detection algorithm and display templates could live in separate files. Not level 1 because sections are clearly labeled and organized rather than an undifferentiated wall of text; not level 3 because there is no one-level-deep reference structure and the file is far larger than the simple-skill threshold. | 2 / 3 |
Total | 10 / 12 Passed |