babysit

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources that could contain malicious instructions designed to manipulate the agent's behavior.
  • Ingestion points: The skill reads pull request comments, reviews, and CI logs via the gh pr view and gh run view commands (SKILL.md).
  • Boundary markers: There are no explicit delimiters used to isolate untrusted input within the agent's context, although the rubric in references/bugbot-triage.md provides a logic for manually triaging findings.
  • Capability inventory: The agent is authorized to modify source code, execute local tests, commit changes, and push to remote repositories (SKILL.md).
  • Sanitization: The skill does not perform automated sanitization of fetched PR content or logs before the agent processes them for code fixes.
  • [COMMAND_EXECUTION]: The skill relies on shell command execution for its primary functionality. It uses the GitHub CLI (gh) to view PR states, fetch logs, and monitor checks, and standard git commands for rebasing, merging, and pushing code changes (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 11:13 PM
Security Audit — agent-trust-hub — babysit