The Agent Skills Directory

[COMMAND_EXECUTION]: The skill involves the construction and execution of shell commands such as git add, git commit, and git switch. This behavior is controlled through a structured workflow that requires the agent to display all proposed actions to the user and wait for explicit confirmation before execution.
[INDIRECT_PROMPT_INJECTION]: The skill reads external project files like CONTRIBUTING.md and AGENTS.md to adopt project-specific commit conventions. This creates a potential surface for indirect prompt injection if those files contain malicious instructions.
Ingestion points: Repository metadata files and documentation (CONTRIBUTING.md, AGENTS.md, docs/).
Boundary markers: No explicit delimiters or warnings are used when interpolating file content into the commit message drafting process.
Capability inventory: The skill has the capability to write to the repository and change branches via the git CLI.
Sanitization: The skill relies on manual user review of the final commit message and file list as the primary safeguard against malicious content ingestion.

git-commit