git-worktree

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses osascript (AppleScript) and tmux to automate terminal operations like creating new tabs, windows, and changing directories. These are standard features for terminal-integration tools and are executed locally within the user's environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes branch and worktree names from the local repository. While this creates a potential surface for indirect prompt injection if an agent processes maliciously named branches, Git's own strict naming conventions (restricting characters like ;, &, |, and backticks) mitigate the risk of shell command injection during interpolation. 1. Ingestion points: git branch, git worktree list. 2. Boundary markers: None present. 3. Capability inventory: git worktree remove, tmux kill-window, osascript (executing terminal commands). 4. Sanitization: The skill relies on Git's internal reference validation.
  • [EXTERNAL_DOWNLOADS]: The skill references official documentation for Git, tmux, and iTerm2. These are trusted external sources and no code or binaries are downloaded from these links.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 06:12 PM