The Agent Skills Directory

[SAFE]: The skill implements a 'Git Safety Protocol' that prevents the execution of destructive commands (e.g., --force, hard reset) and unauthorized configuration changes, ensuring repository integrity.
[SAFE]: Explicit instructions are provided to exclude sensitive files like credentials and private keys from commits, effectively mitigating the risk of accidental data exposure.
[SAFE]: The use of heredocs (cat <<'EOF') for constructing commit messages is a secure pattern that prevents shell injection attacks when handling potentially untrusted input during the commit process.
[INDIRECT_PROMPT_INJECTION]: The skill has a potential attack surface for indirect prompt injection as it ingests untrusted data from local file diffs and status reports via git diff and git status. However, it mitigates these risks by providing clear boundary markers in the form of instructions to focus strictly on semantic analysis for commit messages. The skill's capabilities are restricted to Git operations via the Bash tool, and it implements a 'Git Safety Protocol' alongside specific warnings against processing or committing secret files to sanitize the workflow.

git-commit