The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes standard shell commands including git, gh, echo, and wc to manage the version control workflow. These tools are used for repository status checks, branch management, and commit message validation.\n- [DATA_EXFILTRATION]: Performs network-enabled operations such as git fetch origin and gh repo view to synchronize with and query remote repositories. These operations are standard for git workflows and target the repository's configured origin.\n- [PROMPT_INJECTION]: Features a security guideline that explicitly forbids staging files with sensitive information like .env files, credentials, or private keys, which helps prevent accidental data exposure.\n- [PROMPT_INJECTION]: The skill reads file contents through git diff commands, creating a potential surface for indirect prompt injection from untrusted code changes. \n
Ingestion points: git diff and git diff --staged (SKILL.md)\n
Boundary markers: None present\n
Capability inventory: Shell execution for git, gh, echo, and wc (SKILL.md)\n
Sanitization: None present\n- [COMMAND_EXECUTION]: Uses echo -n "<subject>" | wc -m to verify the character count of the commit subject. This involves interpolation of agent-generated text into a shell command, which is a common functional pattern for length validation.

commit