The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting data from the repository and user arguments to generate commit messages.
Ingestion points: Untrusted data from the repository is fetched using !git diff HEAD and !git log --oneline -10 in SKILL.md, and user input is included via $ARGUMENTS.
Boundary markers: Absent. The skill does not use delimiters or instructions to prevent the agent from following commands that might be found in the diff or commit history.
Capability inventory: The skill can execute shell commands, specifically git add and git commit.
Sanitization: The repository content and user arguments are not sanitized before being processed by the agent.
[COMMAND_EXECUTION]: The skill employs dynamic context injection to execute git status, git diff HEAD, git branch --show-current, and git log --oneline -10 at load time. These commands are legitimate for its purpose. It also instructs the agent to run git add and git commit to perform its primary function.

git-commit