Content
70%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 a strong operational spec: the workflow is clearly sequenced with real validation and edge-case handling, and the parsing logic and output formats are concrete enough to execute directly. Its weaknesses are token efficiency — generic filler sections pad the front — and the complete absence of progressive disclosure, with all templates inlined in one long file.
Suggestions
Delete the generic "Instructions" bullets and the "Use this skill when"/"Do not use this skill when" boilerplate; the pre-flight and data-collection sections already define scope concretely.
Move the two large output-format templates (full project status and single-track status) into a references/ file (e.g., references/output-templates.md) and keep a summary plus the quick/JSON formats in SKILL.md, so the templates load only when needed.
Specify how to populate the "Related Commits" section of the single-track output, e.g. `git log --oneline --grep="{trackId}"`, to close the actionability gap.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The operational sections (pre-flight checks, data collection, calculation logic, output templates) are tight and functional, but the boilerplate "Use this skill when"/"Do not use this skill when" sections and the vague "Instructions" bullets ("Apply relevant best practices and validate outcomes") are filler that add tokens without information — more than the minor trimmable instances of a 4. | 3 / 5 |
Actionability | Guidance is largely executable: exact file paths, counting regexes (/^- \[x\] Task/), a progress-bar formula, complete output templates, and ready-made error messages. The one notable gap is the "Related Commits" section of the single-track output, which never specifies how to retrieve the commits (no git log command or filter), keeping it just short of fully copy-paste ready. | 4 / 5 |
Workflow Clarity | The sequence is explicit — Pre-flight Checks, Data Collection steps 1–4, then Output — with genuine validation checkpoints: initialization verification, three fully specified error states, missing files reported as "unknown", zero-division handled as "N/A", and a read-only guard against mutating state. This is a read-only display skill, so the destructive/batch validation cap does not apply. | 5 / 5 |
Progressive Disclosure | Everything lives in a single 337-line SKILL.md with no references, scripts, or assets at all — roughly 130 lines of output templates plus error-state templates are always inlined in context. Section headers make it navigable, which lifts it above a 2, but there is no progressive disclosure mechanism and template content that could live in a reference file is inline, matching the anchor for content that should be separate. | 3 / 5 |
Total | 15 / 20 Passed |