The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill requests broad execution permissions for the git and gh (GitHub CLI) binaries. This is consistent with its stated purpose of rewriting branch history and creating pull requests.
[PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection (Category 8). An attacker who can influence the content of the Git branch (e.g., via commit messages or file content in a diff) could embed instructions that the agent might follow during the 'Analyze the diff' or 'Reimplement the work' stages.
Ingestion points: The skill reads untrusted data from the local repository via git log and git diff in Step 1 and Step 2.
Boundary markers: Absent. The instructions do not provide delimiters or specific warnings to the model to ignore instructions found within the diffs.
Capability inventory: The skill has the capability to run arbitrary Git commands (git:*), write files, and create pull requests (gh pr create:*).
Sanitization: Absent. There is no evidence of filtering or escaping the data retrieved from the Git history before it is processed by the model.
[SAFE] (SAFE): The use of the --no-verify flag with git commit is explicitly documented as a way to bypass pre-commit hooks for intermediate, potentially broken development steps, and is verified in the final step without the flag.

clean-history