Content
61%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 with copy-paste gh commands and a clear branched workflow, but it is notably verbose due to repeating the full notice text six times and restating title-replacement logic, and it lacks explicit validation checkpoints for its destructive/batch edits.
Suggestions
Define the code-of-conduct notice once (e.g., in a quoted block or a referenced snippet) and reference it from the title, body, and comment sections instead of reproducing the full paragraph six times — this would materially improve conciseness.
Add an explicit verification checkpoint after each destructive edit (e.g., re-fetch the edited issue/comment to confirm the change applied and that all technical content survived sanitization) to lift workflow_clarity above the destructive/batch cap of 3.
Collapse the Notes and Important Principles sections into the Actions steps they restate, removing the duplicated title-replacement explanation to tighten the document.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The identical multi-paragraph code-of-conduct notice is reproduced six times (markdown block plus bash block across the title, body, and comment sections), and the Notes/Important Principles sections restate the title-replacement logic already given in Actions — noticeably verbose with several padded, redundant blocks, fitting the 2 anchor rather than the tighter 3. | 2 / 5 |
Actionability | Provides fully executable, copy-paste-ready gh commands (gh issue view --comments --json, gh issue edit --title/--body, gh issue comment, gh api --method PATCH) covering the common cases with concrete variable interpolation, matching the 5 anchor; placeholders like SANITIZED_BODY_HERE are appropriate for dynamic content. | 5 / 5 |
Workflow Clarity | The Process and Actions sections give a clear sequence with distinct branches and a summary report, but this is a destructive/batch operation (editing issues and replacing comments in place, iterating all comments) with no explicit post-edit verification checkpoint, so the destructive/batch cap of 3 applies rather than 4. | 3 / 5 |
Progressive Disclosure | No bundle files exist and the skill is appropriately self-contained, organized into clearly headed sections (Input Variables, Task, Evaluation, Process, Actions, Important Principles, Notes); the only gap is the long repeated notice that could be factored, fitting the 4 anchor rather than a fully clean 5. | 4 / 5 |
Total | 14 / 20 Passed |