The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard Git CLI commands to manage code changes.
The agent is instructed to use git status, git diff, and git log to analyze repository state.
It performs write operations via git add and git commit to modify the local repository.
In interactive mode, it performs network operations via git push to upload changes to remote servers, though this requires explicit user confirmation via the AskUserQuestion tool.
[INDIRECT_PROMPT_INJECTION]: The skill reads and interprets the content of code changes to generate commit messages, which presents a surface for indirect prompt injection.
Ingestion points: Data enters the agent context through the output of git status, git diff, and git log as described in SKILL.md.
Boundary markers: The skill does not define specific delimiters for separating the untrusted code content from its own instructions.
Capability inventory: The skill has the capability to modify the file system (git add/commit) and access the network (git push).
Sanitization: No automated sanitization is performed, but the instructions explicitly direct the agent to manually review the changes for sensitive content (credentials, secrets, or .env files) and halt the process if found.

git-commit