Content
77%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.
A strong, highly actionable review playbook with an exemplary workflow (explicit checkpoints, feedback loops, severity rubric, and copy-paste gh commands). Its main weaknesses are a monolithic structure that inlines PR-submission mechanics and pentest vectors instead of splitting them into reference files, and repeated badge-image boilerplate that inflates token cost.
Suggestions
Split the PR Review Submission section (gh api reaction/review commands and payload rules) and the Pentest Checklist into references/ files (e.g. references/pr-submission.md, references/pentest-vectors.md), keeping SKILL.md as a lean overview with one-level-deep, clearly signaled links.
Define each priority badge markdown string once (e.g. in a short table or a single reference) and refer to it by priority level elsewhere, instead of repeating the full shields.io URL six times.
Provide concrete substitutions or shell-variable capture for the {owner}/{repo} and <head commit SHA> placeholders in the gh api commands (e.g. derive them from gh pr view --json url,headRefOid) so the submission flow is fully copy-paste ready.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense, expert-level guidance with no padding or explanation of concepts Claude already knows, but the full shields.io badge markdown is repeated verbatim about six times across the severity table, example, and output template, and the security-warning block could be tightened. Efficient with minor trimmable repetition — anchor 4, not 5. | 4 / 5 |
Actionability | Concrete executable commands throughout (gh pr diff, docker run debian:bookworm-slim bash -c, full gh api JSON payloads with --jq filters), exact paths to search, and decision-rule tables. Falls short of anchor 5 only because literal {owner}/{repo} and <head commit SHA> placeholders require manual substitution before the commands run. | 4 / 5 |
Workflow Clarity | A clearly sequenced workflow (scope -> read -> review dimensions -> findings -> output -> PR submission) with explicit validation checkpoints ("If no changes are found, inform the user and stop"; "verify they pass against bash") and feedback loops (retry on invalid line position; self-review forces COMMENT). This matches anchor 5's explicit validation steps, error-recovery loops, and checklists. | 5 / 5 |
Progressive Disclosure | Internally well-sectioned and navigable, but it is a monolithic ~350-line SKILL.md with no bundle files: the PR-submission API mechanics and the pentest checklist are inline detail that belongs in one-level-deep reference files. This fits anchor 3 (content that should be separate is inline) better than anchor 4's "minor organization gaps". | 3 / 5 |
Total | 16 / 20 Passed |