The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill executes git commands using subprocess.run. It correctly implements best practices by passing arguments as a list rather than a single string and avoids the shell=True parameter, which prevents shell injection vulnerabilities.
[DATA_EXPOSURE] (SAFE): The skill maintains a local state file at .claude/worktree-log.json. Analysis shows no attempts to access sensitive system files (e.g., SSH keys, credentials) or exfiltrate data to external servers.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill accepts user-provided names for worktrees which are used in file paths and git branch names.
Ingestion points: The name parameter in create_worktree.
Boundary markers: None present.
Capability inventory: subprocess.run for git operations and Path.mkdir for directory creation.
Sanitization: The use of Python's pathlib for path construction and list-based subprocess arguments provides inherent protection against basic shell injection and directory traversal, though explicit validation of the name string is not present.

ac-workspace-manager