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 with concrete commands, explicit validation gates, and a well-sequenced multi-phase workflow. Its weaknesses are length and repetition that could be tightened, and a monolithic structure that could offload some detail into reference files.
Suggestions
Tighten repeated guidance: state 'form your own opinion but ask the user first' once and reference it, rather than restating it in Phases 4, 5, 6, and the Behavior Rules.
Move reusable lettered-option templates and the People-profiling command patterns into a references/ file, keeping SKILL.md as an overview with one-level-deep links.
Trim the verbose example dialogue blocks to a minimal schema, since the phrasing is largely templated across phases.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and explicitly discourages fluff, and it assumes Claude knows git/gh basics, but at ~240 lines it repeats principles (e.g. 'form your own opinion first') across phases and includes verbose lettered-option templates that could be tightened. | 2 / 3 |
Actionability | Provides copy-paste-ready `gh` commands with exact `--json`/`--jq` flags, concrete git commands, specific file paths, and ready-to-use prompt templates, fully matching the executable-guidance anchor. | 3 / 3 |
Workflow Clarity | An explicit 8-phase sequence with pacing rules, validation gates ('Issue quality gate', 'Understanding Quality Gate'), and a feedback loop in Phase 5 that repeats until the cause is agreed. | 3 / 3 |
Progressive Disclosure | Well organized into clear phases, but it is a single monolithic ~240-line file with no bundle references; content like the People-profiling detail and reusable prompt templates is inline that could be split out. | 2 / 3 |
Total | 10 / 12 Passed |