The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing external data from git diffs. Ingestion points: Data returned by git diff and gh pr diff. Boundary markers: The instructions do not define clear delimiters or warnings to ignore instructions within the diff content. Capability inventory: The skill utilizes shell command execution (git, gh) and multi-turn conversational capabilities. Sanitization: No sanitization or filtering of the diff content is specified before the agent processes and explains it.
[COMMAND_EXECUTION]: The skill executes shell commands using parameters derived from user input or repository state. Evidence: git diff HEAD -- <file>, gh pr diff <number>, and git merge-base commands. While these tools are essential for the skill's functionality, dynamic construction of shell commands with user-provided arguments like branch names or PR numbers presents a potential command injection surface if not handled by a secure tool-calling layer.

walkthrough