The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes dynamic context injection (e.g., !git status, !git log) to automatically gather repository information when the skill is loaded. These commands are standard for development workflows and do not involve remote network calls or sensitive system file access.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands such as git push and gh pr create to manage branches and pull requests.
[PROMPT_INJECTION]: An indirect prompt injection surface is present because the skill ingests external data from git logs and diffs into the agent's context. This data could potentially contain malicious instructions embedded by external contributors in commit messages or code comments.
Ingestion points: SKILL.md (via !git log and !git diff output placeholders)
Boundary markers: Data is wrapped in XML-style tags (e.g., <full_diff>, <commits_on_branch>) which help separate data from instructions
Capability inventory: The skill can execute git and GitHub CLI commands through the Bash tool
Sanitization: No explicit sanitization or instructions to ignore embedded commands are present in the processing logic

pr