@tgoodington/intuition 9.2.2 → 9.3.1

package/docs/archive/TESTING_README.md

@@ -0,0 +1,215 @@
+# Testing Resources for Waldo v3
+
+This directory now contains comprehensive documentation for understanding and testing Waldo v3 (version 3.0.0) - a major architectural upgrade from v2.
+
+## Quick Navigation
+
+### For Understanding the Architecture
+1. **TESTING_SUMMARY.md** - START HERE
+   - Executive summary of v3 changes
+   - Current project state
+   - Major architectural changes
+   - Five-skill orchestration
+   - Testing scope and recommendations
+   - 10+ pages of comprehensive context
+
+2. **SKILL_INTERACTION_GUIDE.md** - How Skills Work Together
+   - Detailed spec for each of the 5 skills
+   - Memory file ownership chart
+   - Data flow diagrams
+   - Common workflows
+   - Troubleshooting guide
+   - File sizes and performance notes
+
+3. **QUICK_TEST_CHECKLIST.md** - For Running Tests
+   - Pre-testing validation
+   - 11 focused test sessions
+   - Edge case testing
+   - Success criteria
+   - Test report template
+
+### Original Documentation
+- **WALDO_V3_COMPLETE_DOCUMENTATION.md** - Full v3 implementation guide
+- **WALDO_V3_DESIGN_SUMMARY.md** - Design philosophy and rationale
+- **docs/intuition-architecture.md** - System design overview
+- **docs/intuition-workflow.md** - Workflow explanation
+- **docs/PROJECT_CONTEXT.md** - Current v2 context (being superseded)
+
+### Skill References
+- **skills/intuition-discovery/references/waldo_core.md** - Waldo v3 implementation
+- **skills/intuition-handoff/references/handoff_core.md** - Handoff orchestration
+- **skills/intuition-outline/references/magellan_core.md** - Planning logic
+- **skills/intuition-execute/references/faraday_core.md** - Execution logic
+
+## Testing Path
+
+### Phase 1: Understanding (2-3 hours)
+1. Read TESTING_SUMMARY.md (Executive Summary through Part 5)
+2. Review SKILL_INTERACTION_GUIDE.md (The Five Skills at a Glance)
+3. Understand the dual dialogue modes and user profile system
+
+### Phase 2: Preparation (1 hour)
+1. Review QUICK_TEST_CHECKLIST.md "Pre-Testing Validation"
+2. Verify all skills are installed
+3. Confirm v3.0.0 in package.json
+4. Set up test project directory
+
+### Phase 3: Testing (4-6 hours depending on depth)
+1. Run Test Session 1-4 (Core functionality)
+2. Run Test Session 5 (Skill coordination)
+3. Run Test Session 6-11 (Specialized testing)
+4. Document findings in test report
+
+### Phase 4: Reporting (1-2 hours)
+1. Compile results using template in QUICK_TEST_CHECKLIST.md
+2. Categorize issues by severity
+3. Make go/no-go decision
+4. Document recommendations for production
+
+## What Changed in v3
+
+### Major Changes
+1. **Waldo Redesign** - From interrogator to research-informed thinking partner
+2. **Dual Dialogue Modes** - Guided (structured) and Open-Ended (natural)
+3. **User Profile System** - Persistent cross-project profile
+4. **Handoff Orchestration** - Explicit phase coordinator
+5. **Profile Integration** - All skills read profile for personalization
+
+### Files Created (v3)
+- `.claude/USER_PROFILE.json` - Global persistent user profile
+- `.claude/USER_PROFILE_template.json` - Template
+- `WALDO_V3_COMPLETE_DOCUMENTATION.md` - Full implementation guide
+- `WALDO_V3_DESIGN_SUMMARY.md` - Design philosophy
+- Testing documentation (this file + others)
+
+### Files Modified (v3)
+- `skills/intuition-discovery/references/waldo_core.md` - Complete redesign
+- `skills/intuition-discovery/SKILL.md` - Updated description
+- `skills/intuition-handoff/references/handoff_core.md` - Profile integration
+- `skills/intuition-outline/references/magellan_core.md` - Profile reading
+- `skills/intuition-execute/references/faraday_core.md` - Profile reading
+- `package.json` - Version 3.0.0
+
+## Key Architectural Concepts
+
+### Research-Informed Thinking Partnership
+Waldo now launches research agents immediately, then engages as a knowledgeable peer who brings insights to dialogue.
+
+### Dual Dialogue Modes
+- **Guided:** Structured options (AskUserQuestion)
+- **Open-Ended:** Natural conversational flow
+- Both produce identical discovery outcomes
+- User choice determines dialogue style only
+
+### Persistent User Profile
+- Located in `.claude/USER_PROFILE.json` (global)
+- Naturally discovered through dialogue
+- Merged by Handoff after each discovery
+- Read by Magellan and Faraday for personalization
+- Tracks role, expertise, communication style, motivation, etc.
+- Includes confidence scores for each property
+- Tracks which projects informed each property
+
+### File-Based Handoffs
+All skill communication happens through files (not APIs):
+- Discovery → `discovery_brief.md` + `discovery_output.json`
+- Handoff → `outline_brief.md` or `execution_brief.md`
+- Outlining → `outline.md`
+- Execution → Implemented code + updated memory
+
+### Five-Skill Orchestration
+```
+/intuition-start → Load context, suggest next step
+  ↓
+/intuition-discovery → Research + dialogue, create discovery outputs
+  ↓
+/intuition-handoff → Extract insights, update memory, merge profile, create outline brief
+  ↓
+/intuition-outline → Synthesize outline, personalize based on profile
+  ↓
+/intuition-handoff → Prepare execution brief
+  ↓
+/intuition-execute → Implement, update memory, personalize based on profile
+```
+
+## Success Criteria
+
+### Functional
+- ✅ Waldo launches research agents in parallel
+- ✅ Dual modes produce equivalent outcomes with different style
+- ✅ User profile discovered naturally and merged accurately
+- ✅ Handoff detects transitions and updates memory correctly
+- ✅ All five skills coordinate via file-based handoffs
+- ✅ Resume capability works across all phases
+
+### Quality
+- ✅ Waldo feels like thinking partner, not interrogator
+- ✅ "Yes, and..." building is evident
+- ✅ Gentle steering is respectful
+- ✅ 1-2 questions per turn feels natural
+- ✅ Cross-sector capability works
+- ✅ Personalization from profile is noticeable
+
+### Architecture
+- ✅ No breaking changes to v2 workflows
+- ✅ File-based architecture maintained
+- ✅ Tool-agnostic (works across Claude Code, Cursor, Copilot)
+- ✅ Security review enforcement maintained
+
+## Important Notes for Testers
+
+1. **Always Check State:** Review `.project-memory-state.json` after each phase - it's the source of truth for workflow status
+
+2. **Profile is Cross-Project:** `.claude/USER_PROFILE.json` is NOT in docs/project_notes/ - it's global
+
+3. **Handoff is Critical:** Never skip handoff between phases. It updates memory and merges profile.
+
+4. **Research is Parallel:** All 3 research agents should launch simultaneously in Waldo's first substantive message
+
+5. **Dialogue Modes are Equivalent:** Both modes should explore same GAPP dimensions with same depth - only style differs
+
+6. **Confidence Scoring Matters:** Profile merge logic uses confidence scores to avoid overwriting with uncertain data
+
+7. **Resume from Files:** All resume capability comes from reading state.json - test this explicitly
+
+## Getting Help During Testing
+
+If you encounter issues:
+
+1. **Check SKILL_INTERACTION_GUIDE.md "Troubleshooting Guide"** - covers common issues
+2. **Review relevant skill's core.md** - detailed implementation patterns
+3. **Check state.json** - often reveals the actual issue
+4. **Consult QUICK_TEST_CHECKLIST.md** - may have a similar test scenario
+
+## Version Info
+
+- **Current Version:** 3.0.0
+- **Latest Commit:** bbfa9ff - "feat: Waldo v3 - Research-informed thinking partnership..."
+- **Major Changes:** Dual dialogue modes, persistent user profile, handoff orchestration
+- **Status:** Production Ready - Ready for Testing
+
+## Documents at a Glance
+
+| Document | Purpose | Length | When to Read |
+|----------|---------|--------|------------|
+| **TESTING_SUMMARY.md** | Comprehensive project overview | 20+ pages | First - understanding |
+| **SKILL_INTERACTION_GUIDE.md** | How skills work together | 15+ pages | Before testing |
+| **QUICK_TEST_CHECKLIST.md** | Test execution guide | 10+ pages | During testing |
+| **WALDO_V3_COMPLETE_DOCUMENTATION.md** | Full implementation guide | 30+ pages | Deep dives |
+| **WALDO_V3_DESIGN_SUMMARY.md** | Design philosophy | 15+ pages | Understanding rationale |
+
+## Next Steps
+
+1. **Read TESTING_SUMMARY.md** to understand what changed and why
+2. **Read SKILL_INTERACTION_GUIDE.md** to understand how skills work together
+3. **Use QUICK_TEST_CHECKLIST.md** to systematically test functionality
+4. **Document findings** using the test report template
+5. **Make go/no-go decision** based on success criteria
+
+---
+
+**Ready to test?** Start with TESTING_SUMMARY.md. It's the complete context you need.
+
+**Questions?** Consult the skill-specific core.md files and the troubleshooting guide in SKILL_INTERACTION_GUIDE.md.
+
+**Let's ship it!**