AI Agentic Harness Guide

Mocking GUI is developed with high quality and consistency through an AI Agentic Harness system. All contributors can leverage specialized agents and workflows to contribute to the project more easily and systematically.

🎯 Mission Map: Workflow Selection Guide

Choose the agent mission that matches your contribution goal.

Case 1: Improve or Add Features to Mocking GUI

• When: You want to modify the GUI panel UI or add new functionality to the core engine.
• How: Run /harness-dev-pipeline in your AI assistant.
• Result: The harness-dev-pipeline workflow orchestrates four sequential phases — Product Planning → Architecture Design → Implementation → Quality Verification — with each phase delegated to a specialized sub-agent. Small fixes get a lightweight single-file run instead of the full artifact set, so the process scales down to the size of the change.

Case 2: Generate Service Mocking Handlers

• When: You want to build high-quality mocking handlers and data based on an API specification.
• How: @handler-specialist, generate handlers based on [Swagger URL or API spec].
• Result: A complete handler set following the 4-Layer architecture (Handlers, Constants, Factories, Utils) is generated in compliance with the handler-specification standard.

Case 3: Project Onboarding and Migration

• When: You want to integrate Mocking GUI into an existing project, or refactor an existing MSW setup to the Mocking GUI standard.
• How: Run /technical-solution-support in your AI assistant.
• Result: The technical-solution-support workflow analyzes your project environment and provides an optimized migration strategy and onboarding guide.

🚀 Agentic Quality Gate

Every workflow automatically performs the following to reduce contributor burden:

1. Automated Test & Build: Runs vitest and build checks to ensure changes don't break existing functionality.
2. Technical Standards Enforcement: Validates real-time compliance with technical-guardrail rules and architecture standards.
3. Documentation Automation: Keeps technical docs and reports in sync with code changes.

Contributors only need to give the agent a clear mission and review and approve the decision report generated at each phase — every artifact (spec, approval gate, ADR) is saved under agent-artifacts/ so it can be reviewed alongside the code change.