The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to automate the review workflow and manage repository state.
Executes git rev-parse HEAD to capture the current commit SHA for audit purposes.
Uses git status --porcelain and git fetch to ensure the local workspace matches the remote state and is free of uncommitted changes.
Utilizes the GitHub CLI (gh) for PR operations, including gh pr view to inspect metadata and gh pr comment to post results.
[EXTERNAL_DOWNLOADS]: The skill interacts with external services to retrieve and update review metadata.
Fetches PR details, state, and previous comments from GitHub's official services via the gh tool.
[PROMPT_INJECTION]: As a review tool, the skill is designed to process and analyze untrusted content from code changes and PR discussions.
Ingestion points: Reads the branch diff (via git diff), the file content of the code under review, and external PR comments from GitHub.
Boundary markers: Uses structured findings tables and explicit role-based prompt templates (e.g., R1, R2, R3) to provide clear context to its sub-reviewers.
Capability inventory: The skill can execute git and gh CLI commands and write findings to local markdown files (review-findings.md).
Sanitization: Employs the --body-file flag when calling gh pr comment, which ensures that the comment body is treated as data rather than shell arguments, mitigating command injection risks.

review-change