review-a-pr-and-report
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill is designed to dynamically discover and execute shell commands specified within the repository under review.
- Evidence: The instructions direct the agent to "Find the gating command" in repository files such as
AGENTS.md, CI workflows, or build files (e.g.,prepublishOnly) and then "Run it in the review checkout." - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content that could contain instructions intended to influence the agent's behavior or verdict.
- Ingestion points: The agent ingests data from
gh pr view(PR body/comments),gh issue view(acceptance criteria), and the repository's own source code and documentation. - Boundary markers: The skill establishes a logical hierarchy of authority (Spec > Ticket > PR Description) but does not implement technical boundary markers or delimiters for the ingested text.
- Capability inventory: The agent has access to
git,ghCLI, and the ability to execute arbitrary shell commands discovered in the repository's configuration. - Sanitization: No explicit sanitization or filtering steps are defined for the content retrieved from external GitHub objects or repository files.
- [COMMAND_EXECUTION]: The skill uses standard development tools to interact with the environment and the repository.
- Evidence: It utilizes
git(diff, log, merge-base) andgh(pr view, pr checks, issue view) to gather context and perform the review.
Audit Metadata