prepare-branch-context

Pass

Audited by Gen Agent Trust Hub on Apr 24, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes git and gh CLI commands to analyze branch deltas, commit history, and pull request metadata. This is standard behavior for repository management tools.
  • [EXTERNAL_DOWNLOADS]: Fetches data from the remote repository origin using git fetch. This operation targets a well-known service (GitHub) to resolve comparison references.
  • [PROMPT_INJECTION]: The skill processes untrusted data from pull request titles, bodies, and commit messages.
  • Ingestion points: Pull request metadata (gh pr view) and commit logs (git log) defined in SKILL.md.
  • Boundary markers: No explicit delimiters are used to separate untrusted content from the summarized output.
  • Capability inventory: The skill is limited to read-only git and gh operations with no file-write or general network capabilities.
  • Sanitization: No filtering is performed on external content. The risk is minimal as the skill only generates a summary for human or agent review.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 24, 2026, 01:19 AM