gh-draft-pr

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill orchestrates local git and gh CLI commands to extract repository information and create pull requests. It generates shell commands incorporating local repository data.
  • [PROMPT_INJECTION]: Indirect prompt injection surface identified through ingestion of untrusted repository data.
  • Ingestion points: Local git commit messages, branch names, and pull request templates.
  • Boundary markers: Absent; the skill does not use specific delimiters to isolate untrusted data from instructions.
  • Capability inventory: Execution of subprocesses (git, gh) and writing temporary files to the local repository.
  • Sanitization: No evidence of escaping or validating branch names or commit messages before they are interpolated into the final PR body or CLI command.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 08:08 PM