Skills
Navigate
skills/gitbobobo/bstack/git-prefs/Gen Agent Trust Hub

git-prefs

Pass

Audited by Gen Agent Trust Hub on Nov 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses git commands and local Node.js scripts to manage the repository state.
  • Evidence: In SKILL.md, the agent is instructed to run node scripts/check.mjs and node <skill_dir>/scripts/empty-worktree.mjs.
  • Evidence: scripts/empty-worktree.mjs utilizes execFileSync to perform git rev-parse and git worktree prune operations.
  • [INDIRECT_PROMPT_INJECTION]: The skill evaluates untrusted repository state to perform automated actions.
  • Ingestion points: SKILL.md describes using git status --porcelain to decide on commit or cleanup actions in files.
  • Boundary markers: The instructions mandate user confirmation for deleting files or modifying .gitignore.
  • Capability inventory: File system modification via rmSync (in scripts/empty-worktree.mjs) and git commits/syncing.
  • Sanitization: The empty-worktree.mjs script verifies that the current directory is a linked worktree and not the main repository before allowing deletion.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 8, 2026, 09:19 AM