pr-refiner
Pass
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection due to the processing of untrusted data within the agent's context.
- Ingestion points: Processes external data from GitHub PR titles, descriptions, file contents, and review threads (especially Codex/AI comments mentioned in
github-graphql.md). - Boundary markers: Absent. The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when the agent reads PR comments or file diffs.
- Capability inventory: Performs local file edits,
git commit,git push, and executes GitHub GraphQL mutations to resolve review threads (gh api graphql). - Sanitization: Absent. There is no explicit instruction to sanitize or escape the content of PR comments before they influence the agent's refinement logic.
- [COMMAND_EXECUTION]: Executes shell commands via the GitHub CLI (
gh) and Git to manage PR state and resolve discussions. - Evidence: Uses
gh api graphqlingithub-graphql.mdto fetch and resolvePullRequestReviewThreadnodes. This is the core functionality intended by the skill.
Audit Metadata