cursor-ai-toolkit 1.0.0

package/commands/context/context-prune.md

@@ -0,0 +1,201 @@
+---
+description: Summarize session progress, clean context, prevent AI drift in long chats
+category: Context & Analysis
+aliases: [prune, summarize, cleanup]
+---
+
+# Context Prune - Manage Context Window & Prevent Drift
+
+Clean chat history and summarize progress to maintain AI performance.
+
+## Usage
+
+```
+/context-prune
+/context-prune --summarize              # Save progress summary
+/context-prune --fresh                  # Start clean with context
+/context-prune --checkpoint {NAME}      # Named checkpoint
+```
+
+## Why This Matters
+
+In 2025, **context window management is the #1 performance bottleneck**:
+
+-   Long conversations cause "context drift"
+-   AI forgets earlier decisions
+-   Response quality degrades over time
+-   Hallucinations increase
+
+## What This Does
+
+1. **Analyzes current context** - Measures token usage, key decisions
+2. **Creates checkpoint** - Saves essential state
+3. **Prunes irrelevant** - Removes debug output, repeated content
+4. **Generates summary** - Compressed handoff for fresh session
+5. **Restores critical** - Ensures key context is preserved
+
+## Context Health Indicators
+
+| Indicator          | Healthy | Warning   | Critical  |
+| ------------------ | ------- | --------- | --------- |
+| Token usage        | < 50%   | 50-80%    | > 80%     |
+| Conversation turns | < 20    | 20-40     | > 40      |
+| Files discussed    | < 15    | 15-30     | > 30      |
+| Decisions made     | Tracked | Some lost | Forgotten |
+
+## Output Format
+
+````
+📋 Context Analysis...
+
+════════════════════════════════════════════════════════════════
+  CONTEXT HEALTH
+════════════════════════════════════════════════════════════════
+
+## Current State
+
+| Metric | Value | Status |
+|--------|-------|--------|
+| Estimated tokens | 45,000 | 🟡 56% used |
+| Conversation turns | 28 | 🟡 Warning |
+| Files discussed | 12 | ✅ Healthy |
+| Tools invoked | 67 | Normal |
+
+## Drift Detection
+
+⚠️ Potential drift detected:
+- Turn 5: Decided to use `useWatch` over `watch()`
+- Turn 23: Used `watch()` (inconsistent)
+
+⚠️ Forgotten context:
+- Original goal: "Implement TICKET-123"
+- Current focus: Debugging unrelated styling
+
+════════════════════════════════════════════════════════════════
+  CHECKPOINT CREATED
+════════════════════════════════════════════════════════════════
+
+## Session Summary (for handoff)
+
+### Original Goal
+Implement marketing texts for protection packages (TICKET-123)
+
+### Key Decisions Made
+1. Use `display_name` from backend with Statsig fallback
+2. Prevent duplicate titles using `useOXBreakpoint`
+3. Keep `originalName` for modal title preservation
+
+### Files Modified
+- LineItems.tsx - Added conditional TooltipTitle
+- ProtectionPackageCard.styled.ts - Added TooltipTitle component
+- helpers.ts - Updated getModifiedLineItems logic
+- PackagesV2.tsx - Removed duplicate function
+
+### Current State
+- [x] Implementation complete
+- [x] Tests passing
+- [ ] PR not created
+- [ ] How to Test not added
+
+### Next Steps
+1. Create PR with `/jira-pr TICKET-123`
+2. Add How to Test with `/jira-test TICKET-123`
+
+### Context to Preserve
+```typescript
+// Key pattern established
+const { biggerThanSmall } = useOXBreakpoint();
+{biggerThanSmall && originalName && <TooltipTitle ... />}
+````
+
+════════════════════════════════════════════════════════════════
+RECOMMENDATIONS
+════════════════════════════════════════════════════════════════
+
+Option A: Continue in current session
+
+-   Risk: Moderate context drift
+-   Token headroom: 44% remaining
+
+Option B: Start fresh session (Recommended)
+
+-   Copy the summary above
+-   Paste at start of new conversation
+-   Full context restoration guaranteed
+
+Checkpoint saved as: `TICKET-123-checkpoint-2024-12-23`
+
+````
+
+## Checkpoint File Format
+
+When `--checkpoint` is used, creates:
+
+```markdown
+# Context Checkpoint: {NAME}
+
+**Created:** {timestamp}
+**Session ID:** {id}
+**Tokens used:** {count}
+
+## Goal
+{original goal}
+
+## Decisions
+{numbered list}
+
+## State
+{current progress}
+
+## Files
+{list of modified files}
+
+## Resume Instructions
+Paste this into a new chat to continue:
+
+---
+{compressed context for handoff}
+---
+````
+
+## Auto-Prune Triggers
+
+Consider running `/context-prune` when:
+
+-   Conversation exceeds 30 turns
+-   AI starts repeating itself
+-   AI forgets earlier decisions
+-   Response quality noticeably drops
+-   Switching to a different task
+
+## Context Preservation Strategies
+
+### Essential (Always Preserve)
+
+-   Original goal/ticket
+-   Key architectural decisions
+-   File paths modified
+-   Patterns established
+
+### Prunable (Safe to Drop)
+
+-   Debug output
+-   Error messages (once resolved)
+-   Exploratory searches
+-   Redundant explanations
+
+### Compressible
+
+-   Long code blocks → File references
+-   Repeated patterns → Single example
+-   Discussion → Decision summary
+
+## AI Execution
+
+When user runs `/context-prune`:
+
+1. **Analyze context** - Token count, turn count, drift indicators
+2. **Identify decisions** - Extract key choices made
+3. **Summarize state** - Current progress and next steps
+4. **Create checkpoint** - Portable summary for handoff
+5. **Recommend action** - Continue or fresh session

package/commands/context/orchestrate.md

@@ -0,0 +1,279 @@
+---
+description: Spawn specialized sub-agents for parallel task execution (multi-agent)
+category: Workflow Orchestration
+aliases: [multi-agent, spawn, coordinate]
+---
+
+# Orchestrate - Multi-Agent Coordination
+
+Spawn specialized sub-agents for parallel task execution.
+
+## Usage
+
+```
+/orchestrate {GOAL}
+/orchestrate --ticket {TICKET_ID}
+/orchestrate --pr {PR_NUMBER} --mode review
+/orchestrate --feature "Add loyalty program"
+```
+
+## What This Does
+
+1. **Parses goal** - Breaks down into specialized tasks
+2. **Spawns agents** - Each agent handles specific domain
+3. **Parallel execution** - Tasks run simultaneously
+4. **Coordination** - Merges results, resolves conflicts
+5. **Final synthesis** - Unified output
+
+## Available Agents
+
+| Agent                | Role               | Specialization                   |
+| -------------------- | ------------------ | -------------------------------- |
+| 🏗️ **Architect**     | System design      | Validates patterns, dependencies |
+| 🔒 **Security**      | Vulnerability scan | Secrets, SAST, dependencies      |
+| 📝 **Documentation** | Docs update        | README, Confluence, JSDoc        |
+| 🧪 **Testing**       | Test generation    | Unit, integration, E2E           |
+| 🎨 **Styling**       | Design compliance  | Token usage, responsive          |
+| 📊 **Analytics**     | Tracking           | Events, metrics, logging         |
+| ⚡ **Performance**   | Optimization       | Bundle, runtime, memory          |
+
+## Orchestration Flow
+
+```
+┌─────────────────────────────────────────────────────────────┐
+│  /orchestrate "Implement TICKET-123"                          │
+└─────────────────────────────────────────────────────────────┘
+                            ↓
+┌─────────────────────────────────────────────────────────────┐
+│  GOAL DECOMPOSITION                                         │
+├─────────────────────────────────────────────────────────────┤
+│  1. Understand requirements (Architect)                     │
+│  2. Implement feature (Main Agent)                          │
+│  3. Check security (Security Agent)                         │
+│  4. Generate tests (Testing Agent)                          │
+│  5. Update docs (Documentation Agent)                       │
+│  6. Verify styling (Styling Agent)                          │
+└─────────────────────────────────────────────────────────────┘
+                            ↓
+┌─────────────────────────────────────────────────────────────┐
+│  PARALLEL EXECUTION                                         │
+├─────────────────────────────────────────────────────────────┤
+│                                                             │
+│  ┌──────────┐  ┌──────────┐  ┌──────────┐                   │
+│  │ Security │  │ Testing  │  │   Docs   │                   │
+│  │  Agent   │  │  Agent   │  │  Agent   │                   │
+│  └────┬─────┘  └────┬─────┘  └────┬─────┘                   │
+│       │             │             │                         │
+│  ┌────▼─────────────▼─────────────▼────┐                    │
+│  │         COORDINATION LAYER          │                    │
+│  └─────────────────────────────────────┘                    │
+│                                                             │
+└─────────────────────────────────────────────────────────────┘
+                            ↓
+┌─────────────────────────────────────────────────────────────┐
+│  SYNTHESIS & OUTPUT                                         │
+├─────────────────────────────────────────────────────────────┤
+│  • Merge all agent outputs                                  │
+│  • Resolve any conflicts                                    │
+│  • Generate unified report                                  │
+│  • Apply all changes atomically                             │
+└─────────────────────────────────────────────────────────────┘
+```
+
+## Output Format
+
+````
+📋 Orchestrating: Implement TICKET-123...
+
+════════════════════════════════════════════════════════════════
+  GOAL DECOMPOSITION
+════════════════════════════════════════════════════════════════
+
+Feature: Marketing texts for protection packages
+Ticket: TICKET-123
+Complexity: Medium
+
+Tasks identified:
+1. [MAIN] Implement display_name support in LineItems
+2. [SECURITY] Validate input sanitization for marketing texts
+3. [TESTING] Generate tests for new display logic
+4. [DOCS] Update component documentation
+5. [STYLING] Verify TooltipTitle spacing
+
+════════════════════════════════════════════════════════════════
+  AGENT EXECUTION
+════════════════════════════════════════════════════════════════
+
+## 🏗️ Architect Agent
+
+Status: ✅ Complete
+
+Analysis:
+- Pattern: Consistent with existing tooltip implementation
+- Dependencies: No new external dependencies
+- Breaking changes: None
+- Recommendation: ✅ Proceed with implementation
+
+## 🔒 Security Agent
+
+Status: ✅ Complete
+
+Scanned:
+- [ ] Hardcoded secrets: None found
+- [ ] XSS vectors: `display_name` passed through OXRichTextBlock (sanitized) ✅
+- [ ] Injection risks: None detected
+- [ ] Dependency vulnerabilities: None new
+
+Result: ✅ No security issues
+
+## 🧪 Testing Agent
+
+Status: ✅ Complete
+
+Generated:
+```typescript
+// LineItems.test.tsx
+describe('LineItems', () => {
+    it('should render display_name when available', () => { ... });
+    it('should fallback to name when display_name missing', () => { ... });
+    it('should show title only on desktop', () => { ... });
+});
+````
+
+Files created:
+
+-   LineItems.test.tsx (3 test cases)
+-   helpers.test.ts (5 test cases)
+
+## 📝 Documentation Agent
+
+Status: ✅ Complete
+
+Updated:
+
+-   [ ] Component JSDoc comments
+-   [ ] Type definitions (ILineItemInfo.originalName)
+-   [ ] README section on marketing texts
+
+Confluence page created:
+
+-   "Protection Package Display Names" (875167748)
+
+## 🎨 Styling Agent
+
+Status: ✅ Complete
+
+Verified:
+
+-   [ ] TooltipTitle uses spacing('xs') ✅
+-   [ ] Responsive breakpoint (useOXBreakpoint) ✅
+-   [ ] No hardcoded values ✅
+-   [ ] Design token compliance ✅
+
+════════════════════════════════════════════════════════════════
+SYNTHESIS
+════════════════════════════════════════════════════════════════
+
+## Changes Summary
+
+| Agent   | Files Modified | Files Created  |
+| ------- | -------------- | -------------- |
+| Main    | 5              | 0              |
+| Testing | 0              | 2              |
+| Docs    | 1              | 1 (Confluence) |
+
+## Conflict Resolution
+
+No conflicts detected between agents.
+
+## Final Checklist
+
+✅ Implementation complete
+✅ Security verified
+✅ Tests generated
+✅ Documentation updated
+✅ Styling validated
+✅ All agents agree
+
+════════════════════════════════════════════════════════════════
+NEXT STEPS
+════════════════════════════════════════════════════════════════
+
+Ready to:
+
+1. Create PR with all changes
+2. Run full test suite
+3. Request review
+
+Run: /jira-pr TICKET-123 to create PR
+
+````
+
+## Agent Configurations
+
+### Architect Agent
+```yaml
+focus:
+  - System design patterns
+  - Dependency analysis
+  - Breaking change detection
+  - Architecture compliance
+tools:
+  - Dependency graph analysis
+  - Pattern matching
+  - CODEOWNERS lookup
+````
+
+### Security Agent
+
+```yaml
+focus:
+    - Secret scanning
+    - XSS/injection detection
+    - Dependency CVEs
+    - Auth/authz checks
+tools:
+    - gitleaks (secrets)
+    - semgrep (SAST)
+    - npm audit (deps)
+```
+
+### Testing Agent
+
+```yaml
+focus:
+    - Unit test generation
+    - Integration tests
+    - Edge case coverage
+    - Mock generation
+tools:
+    - /risk-test-gen
+    - /mutation-audit
+    - Jest templates
+```
+
+### Documentation Agent
+
+```yaml
+focus:
+    - JSDoc updates
+    - README maintenance
+    - Confluence sync
+    - Type documentation
+tools:
+    - /jira-docs
+    - JSDoc generator
+    - Confluence API
+```
+
+## AI Execution
+
+When user runs `/orchestrate {GOAL}`:
+
+1. **Parse goal** - Break into domain-specific tasks
+2. **Assign agents** - Map tasks to specialized agents
+3. **Execute parallel** - Run agents simultaneously
+4. **Collect results** - Gather outputs from each agent
+5. **Resolve conflicts** - Handle overlapping changes
+6. **Synthesize** - Merge into unified output
+7. **Report** - Summary of all agent actions