The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses Python scripts to execute git and gh CLI commands. These are executed via subprocess.run using argument lists rather than shell strings, which is a secure practice that prevents traditional command injection via malicious inputs.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface due to how it handles repository data.
Ingestion points: The scripts/gather_context.py script retrieves commit messages and code diffs from the local repository using git log and git diff commands.
Boundary markers: The instructions in SKILL.md lack explicit boundary markers or instructions for the AI to ignore potentially malicious instructions embedded within the retrieved git history or code changes.
Capability inventory: The skill has the capability to modify repository metadata through the scripts/update_description.py script, which uses gh pr edit to update pull request descriptions.
Sanitization: There is no filtering or sanitization performed on the content retrieved from git before it is presented to the AI agent for summarization.

pr-desc