The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes repository data (file status, diffs, and logs) to generate commit messages and determine how to resolve merge conflicts. This presents an indirect injection surface where malicious content embedded in the codebase could attempt to influence the agent's output.
Ingestion points: Analyzes data from the local repository environment using git status, git diff, git diff --cached, and git log (SKILL.md, Instructions Step 1, 2 and Merge Conflict Protocol Step 3).
Boundary markers: The instructions do not define specific delimiters or "ignore" instructions for the ingested code content.
Capability inventory: The skill can modify the repository state and network state via git commit, git push, and git pull (SKILL.md, Instructions Step 5, 7 and Merge Conflict Protocol).
Sanitization: No automated sanitization is specified, but the skill includes a "Golden Rule" requiring explicit user confirmation before resolving conflicts and a tip to manually check for secrets, which significantly mitigates the risk.
[COMMAND_EXECUTION]: The skill executes several shell commands to manage the git lifecycle.
Evidence: Execution of git status, git diff, git commit, git push, and git pull (SKILL.md).
Safety Note: The skill uses a quoted heredoc ('EOF') in the git commit command pattern, which is a security best practice that prevents shell expansion and reduces the risk of unintended command injection from generated commit messages.

commit