The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes standard shell commands including git and sed to interact with the local filesystem and manage version control state. These operations are essential for the skill's primary function of automating code commits and repository management.
[DATA_EXFILTRATION]: The skill performs network operations via git push origin to send local repository data and commit history to a remote server. This behavior is documented as the skill's intended purpose for sending changes upstream to the configured origin.
[PROMPT_INJECTION]: The skill processes repository metadata and file differences to generate commit messages, creating a surface for indirect prompt injection where malicious content in a file diff could influence the agent's output.
Ingestion points: Git diff output and repository files such as .github/PULL_REQUEST_TEMPLATE.MD (SKILL.md).
Boundary markers: None present in the instructions for processing diff content.
Capability inventory: Shell command execution and remote data transmission via Git (SKILL.md).
Sanitization: None; the agent is directed to generate subjects directly from the diff content.

git-commit-push