learn-from-code-review

Pass

Audited by Gen Agent Trust Hub on Sep 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external GitHub PR review comments, which could contain malicious instructions designed to influence the agent's behavior during distillation or in the resulting generated skills. * Ingestion points: Fetches review comment bodies via gh api calls in SKILL.md (Step 2). * Boundary markers: None identified to separate untrusted data from instructions. * Capability inventory: The skill creates new files in .openhands/skills/, updates AGENTS.md, and uses the create_pr tool to propose changes. * Sanitization: No specific content sanitization or escaping is mentioned before the data is processed or written to files.
  • [COMMAND_EXECUTION]: The skill utilizes the GitHub CLI (gh) to interact with repositories, which is appropriate for its stated purpose but involves executing shell commands with repository-specific arguments.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 27, 2026, 03:37 PM
Security Audit — agent-trust-hub — learn-from-code-review