소개
The Agent Output Formats skill provides a rigorous set of structured templates for the research, planning, implementation, and review phases of autonomous development. By enforcing uniform markdown structures for agent outputs—including research findings, architecture plans, and security audits—it reduces cognitive load for developers and ensures seamless parsing for automated workflows. This skill leverages Claude Code 2.0’s progressive disclosure architecture to significantly reduce prompt token usage while maintaining professional standards for documentation, commit messages, and pull requests.
