git-branching

Pass

Audited by Gen Agent Trust Hub on Oct 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses local shell scripts to automate Git and GitHub CLI (gh) operations. These operations (fetch, rebase, merge, worktree management) are restricted to the local repository and the configured remote.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data from GitHub PR states and Git branch metadata to determine its workflow logic.
  • Ingestion points: Pull request JSON data from gh pr view, Git branch names, and output from helper scripts in SKILL.md.
  • Boundary markers: Helper scripts use structured key-value output formats (e.g., rerun_required=true) to pass status to the agent.
  • Capability inventory: The skill uses git and gh subprocess calls for repository management across all provided scripts.
  • Sanitization: Shell scripts use set -euo pipefail for error handling and include explicit checks to prevent the deletion of the target base branch.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 13, 2026, 04:27 PM