git-structured

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates the execution of Git system commands through a JavaScript wrapper. It supports a wide range of operations including repository status, commit management, and branch control.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to its requirement to process untrusted text from Git repositories.
  • Ingestion points: Data from external repositories is ingested through operations like git log, git diff, and git status (SKILL.md).
  • Boundary markers: The skill returns data in a structured JSON format, which helps the agent distinguish between data and instructions, though it does not fully eliminate the risk of the model obeying embedded commands.
  • Capability inventory: The skill can perform file system writes (commits), network operations (git push and git pull), and repository restructuring (branch management).
  • Sanitization: Documentation indicates the implementation of argument sanitization and the blocking of dangerous flags such as --exec= and --upload-pack= to prevent command injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 05:10 AM