The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses standard local commands such as git, cd, and dirname to facilitate repository management tasks. These operations are restricted to the local environment and the user's repository scope.
[PROMPT_INJECTION]: Indirect surface analysis:
Ingestion points: Data enters the context via git worktree list --porcelain and git show-ref as described in SKILL.md.
Boundary markers: There are no explicit instructions to ignore prompt-like content within git output.
Capability inventory: The skill has the capability to execute git subcommands and navigate the local filesystem.
Sanitization: The skill implements specific sanitization logic for directory 'slugs' to prevent the use of invalid or dangerous characters in the filesystem, although it relies on standard AI processing for repository-provided data such as branch names.

worktree