Summary

Update the RAI Planner agent system to align with the Microsoft Responsible AI Impact Assessment Guide. This includes:

• Align screening and assessment patterns with the Guide's recommended approach
• Add disclaimer and copyright attribution
• Implement UX improvements for smoother session flow
• Add user-customizable output specifications
• Improve phase transitions and confirmation gates

Scope

11 files across .github/agents/rai-planning/, .github/instructions/rai-planning/, and .github/prompts/rai-planning/.

Acceptance Criteria

• All references updated to Impact Assessment Guide and Template
• Disclaimer present in agent and prompt files
• Copyright attribution added
• UX improvements implemented (pre-scan, lighter gates, progressive disclosure, running summary)
• User preferences for output customization working
• All validation passes