git-commit

The git-commit skill provides an AI-driven interface for managing Git repositories, specifically focusing on generating standardized, semantic commit messages. By analyzing the current working directory's diff status, the agent helps developers adhere to the Conventional Commits specification, ensuring a clean and consistent project history. This tool is designed for software engineers who want to automate the boilerplate of writing commit messages while maintaining high-quality documentation for every code change. It serves as an intelligent assistant that handles the technical nuances of Git, allowing developers to focus on feature implementation rather than version control mechanics.

• Automatically detects change types (feat, fix, docs, refactor, perf, etc.) and scopes based on diff content.

• Generates semantic commit messages using imperative mood and present tense, adhering to the 72-character length limit for descriptions.

• Supports intelligent staging of files, allowing for granular grouping of logical changes before committing.

• Handles advanced Git concepts like BREAKING CHANGE markers and issue references (e.g., Closes #123).

• Performs safety checks to prevent common pitfalls such as staging sensitive files (e.g., .env, credentials) or performing destructive operations like force pushes.

• Use this skill when prompted to commit changes, create a new commit, or when mentioning /commit in the chat.

• The agent evaluates both staged and unstaged changes using git diff and git status, providing the most accurate analysis possible.

• Users can manually override the automatically detected type, scope, or description via an interactive prompt.

• The tool is designed with a safety protocol: it will not modify git configuration, perform force-resets, or skip hooks unless explicitly requested by the user.

• It facilitates a structured workflow: 1. Analyze diffs, 2. Stage logical groups of files, 3. Generate semantic messages, and 4. Execute the commit safely.