monitor-pr
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- Command Execution: The skill executes commands using the
gh(GitHub) CLI to view repository details, check PR status, and interact with the GitHub API. These operations are essential for its purpose of monitoring pull request activity and state. - Indirect Prompt Injection Consideration: The skill processes data from external sources, specifically Pull Request comments and CI check outputs. This represents a standard surface for indirect interaction where external content (like review comments) is presented to the agent for processing.
- Ingestion points: Untrusted data enters the agent context via the stdout of
wait-for-ci.mts(CI check names and links) andwait-for-copilot-review.mts(review comment bodies) after being fetched from the GitHub API. - Boundary markers: The output uses structured labels such as
RESULT:,FAILED_CHECKS:, andComments:to help the agent distinguish between script status and external content. - Capability inventory: The skill instructions enable the agent to modify files, resolve GitHub review threads, and rerun CI jobs using
gh run rerun. - Sanitization: The scripts retrieve data directly from the GitHub API using the
ghtool. While it parses the structured JSON, the content of PR comments is passed through as-is for the agent to evaluate. - Automated Workflow Actions: The skill includes logic for retrying CI jobs that are identified as 'flaky'. This automation is a productivity feature that relies on the agent's analysis of failure logs to determine if a rerun is appropriate without manual intervention.
Audit Metadata