review-and-simplify-changes

Pass

Audited by Gen Agent Trust Hub on Jul 12, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from source code and repository metadata which may contain malicious instructions designed to manipulate the agent.
  • Ingestion points: Reads git diff outputs, specific files provided by the user, and local project documentation like AGENTS.md.
  • Boundary markers: There are no explicit delimiters or protective instructions defined to isolate external code content from the agent's logic.
  • Capability inventory: The skill has the ability to modify local files, spawn multiple sub-agents for parallel processing, and execute local shell commands for validation purposes.
  • Sanitization: The skill lacks formal sanitization or validation of the ingested code beyond instructions to skip subjective refactors.
  • [COMMAND_EXECUTION]: The skill invokes local shell commands to identify changes and validate fixes.
  • Evidence: Executes git diff commands and runs project-specific validation tools such as test runners, compilers, and linters.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 12, 2026, 04:58 PM
Security Audit — agent-trust-hub — review-and-simplify-changes