Content
100%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is a tight, executable workflow with concrete gh/git commands, explicit validation checkpoints, and a well-defined feedback loop, organized into clearly labeled sections. It adds only what Claude would not already know and stays within a single appropriately structured file.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes Claude's competence — no filler explaining what a PR or CI check is, and every line is a command, a state classification, or an exit condition. Nothing to trim without losing useful signal. | 3 / 3 |
Actionability | Provides copy-paste-ready executable commands (gh pr view, gh api check-runs/status, gh run view --log-failed, git push) plus a failure-type → action table. Guidance is concrete and directly runnable. | 3 / 3 |
Workflow Clarity | Clear 8-step sequence with explicit validation checkpoints (poll until terminal state, 'Only rerun a failed job once without code changes', 'loop is only complete when the remote PR checks ... are green') and a genuine fix → push → refresh SHA → restart feedback loop. | 3 / 3 |
Progressive Disclosure | No bundle files are present; the single self-contained SKILL.md is well-organized into Scope, Inputs, Workflow, Output, and Notes sections. Per the rubric's simple-skills note, a clear single-file structure scores 3 without external references. | 3 / 3 |
Total | 12 / 12 Passed |