commit-messages

This skill provides a standardized workflow for software developers to create high-quality, descriptive, and consistent git commit messages. By analyzing staged changes using git diff --staged, the agent identifies the nature of modifications—such as features, bug fixes, refactors, or performance improvements—and maps them to the Conventional Commits specification. This ensures that repository history remains clean, readable, and machine-parsable for automated changelog generation and versioning tools.

The skill is designed for developers working in professional environments who need to adhere to strict version control practices without the overhead of manual message drafting. It intelligently suggests appropriate types and scopes based on the analyzed code context, effectively bridging the gap between raw code changes and professional project documentation.

• Automatically analyzes staged git diffs to summarize code changes.

• Enforces the Conventional Commits format, including type, scope, subject line, body, and footer.

• Supports standard commit types including feat, fix, docs, style, refactor, perf, test, build, ci, chore, and revert.

• Adheres to best practices like imperative mood, maximum line length constraints (72 characters), and proper formatting for breaking changes.

• Includes support for issue references, allowing users to link commits directly to tracking systems like GitHub Issues or Jira.

• Provides an interactive feedback loop where the agent shows the summary, proposes the message, and allows the user to review or modify before finalization.

• Usage notes: Always stage your files (git add) before triggering the skill to ensure the agent has full visibility into the changes.

• Ideal for use during the pre-commit phase, when reviewing changes, or while preparing releases.

• Input: Git staged changes; Output: A formatted commit message string ready for execution.

• Practical tips: If the agent's proposed scope is incorrect, provide the correct context in the follow-up prompt to improve accuracy.

• This tool is a core component of the Claude Code Mastery toolkit, emphasizing deterministic quality gates and repository standard enforcement.