mergepatrol

Pass

Audited by Gen Agent Trust Hub on Nov 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and process untrusted data from external sources, which creates a risk of indirect prompt injection.
  • Ingestion points: The agent is directed to read PR comments, unresolved review threads, and findings from Mergepatrol's automated review as specified in the 'What counts as a comment' and 'The loop' sections of SKILL.md.
  • Capability inventory: The agent has the capability to execute GitHub CLI commands (gh pr edit, gh stack), interact with Mergepatrol MCP tools to modify PR states, and is expected to push commits to fix review findings.
  • Boundary markers: The skill lacks explicit boundary markers or instructions to treat comment content as untrusted data, increasing the risk that instructions embedded in comments could influence agent behavior.
  • Sanitization: There are no instructions for sanitizing or validating the content of the comments before the agent acts upon them.
  • [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) and a suite of MCP tools (queue_pull, dequeue_pull, pull_status, list_queue) to manage pull request states and labels. These operations are consistent with the skill's stated purpose of PR management.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 2, 2026, 05:05 PM