cursor-ai-toolkit 1.0.0

package/commands/context/smart-detection.md

@@ -0,0 +1,272 @@
+---
+description: Auto-detect AC changes, multiple PRs needed, optimal workflow decisions
+category: AI Self-Improvement
+aliases: [detect, auto-detect]
+---
+
+# Smart Detection - Intelligent Workflow Decisions
+
+Automatic detection of changes, multiple PRs, and smart decision-making.
+
+## 1. AC Change Detection
+
+When acceptance criteria changes, I detect and update affected steps.
+
+### How It Works
+
+```bash
+# Compare current AC with what was used for implementation
+1. Fetch current Jira AC
+2. Compare with AC stored when implementation started
+3. Identify what changed
+4. Determine affected steps
+5. Suggest updates
+```
+
+### Detection Output
+
+```
+════════════════════════════════════════════════════════════════
+  ⚠️ ACCEPTANCE CRITERIA CHANGED
+════════════════════════════════════════════════════════════════
+
+Original AC (when branch created):
+  ✅ Display name shown when available
+  ✅ Tooltip shows original name as title
+  ✅ Fallback to original name when missing
+
+Current AC (just fetched):
+  ✅ Display name shown when available
+  ✅ Tooltip shows original name as title
+  ✅ Fallback to original name when missing
+  🆕 Display name should be bold (NEW)
+  🆕 Add analytics tracking for display name views (NEW)
+
+════════════════════════════════════════════════════════════════
+  AFFECTED STEPS
+════════════════════════════════════════════════════════════════
+
+The following need updates:
+
+1. 📝 Implementation
+   - Add bold styling to display name
+   - Add analytics tracking
+
+2. 📋 How to Test (already posted)
+   - Needs update to include new AC items
+
+3. 🔍 Self-Review
+   - Re-run with new criteria
+
+Actions:
+  A) Update implementation + How to Test
+  B) Update How to Test only (implementation pending)
+  C) Ignore (AC change not relevant)
+
+Choose: (a/b/c)
+```
+
+### Auto-Update How to Test
+
+```bash
+# When AC changes, I can auto-update the Jira comment
+jira issue comment add {TICKET_ID} '**How to Test (Updated):**
+
+[Previous steps remain valid]
+
+**New Items:**
+5. **Verify Bold Styling:**
+   - Display name should appear in bold
+
+6. **Verify Analytics:**
+   - Open browser dev tools → Network tab
+   - Confirm tracking event fires on display name view
+
+---
+_Updated due to AC changes on {DATE}_'
+```
+
+## 2. Multiple PRs Intelligence
+
+When multiple PRs are attached to a ticket, I determine the main one.
+
+### Detection Logic
+
+```
+Priority order for "main PR":
+1. PR with most recent activity
+2. PR with most files changed
+3. PR created by you
+4. PR with "main" or "primary" in title
+5. PR with most review comments (indicates active review)
+```
+
+### Detection Output
+
+```
+════════════════════════════════════════════════════════════════
+  🔗 MULTIPLE PRs DETECTED FOR TICKET-123
+════════════════════════════════════════════════════════════════
+
+Found 3 PRs linked to this ticket:
+
+| PR | Title | Author | Status | Files | Activity |
+|----|-------|--------|--------|-------|----------|
+| #23043 | [TICKET-123] Display name support | you | Draft | 6 | 2h ago |
+| #23010 | [TICKET-123] Backend integration | backend-dev | Merged | 3 | 3d ago |
+| #22980 | [TICKET-123] Initial prototype | you | Closed | 2 | 1w ago |
+
+════════════════════════════════════════════════════════════════
+  ANALYSIS
+════════════════════════════════════════════════════════════════
+
+🎯 Main PR: #23043 (Display name support)
+   Reason: Most recent activity, authored by you, draft status
+
+📦 Backend PR: #23010 (already merged)
+   Context: Backend changes are complete, frontend depends on this
+
+🗑️ Obsolete: #22980 (closed prototype)
+   Can be ignored
+
+Working with PR #23043. Correct? (y/n)
+```
+
+### Smart Context Gathering
+
+When multiple PRs exist, I gather context from all:
+
+```
+From merged PR #23010 (backend):
+- New field: display_name added to API response
+- Optional field, may be null
+
+From closed PR #22980 (prototype):
+- Initial approach: inline styling
+- Rejected because: didn't follow component patterns
+
+Applied to current PR #23043:
+- Use display_name from backend (#23010)
+- Follow component patterns (learned from #22980)
+```
+
+## 3. PR Size Alerts
+
+When a PR becomes too large, I suggest splitting.
+
+### Thresholds
+
+| Metric         | Warning  | Critical |
+| -------------- | -------- | -------- |
+| Lines changed  | > 200    | > 400    |
+| Files changed  | > 8      | > 15     |
+| Review time    | > 20 min | > 45 min |
+| Teams affected | > 2      | > 3      |
+
+### Alert Output
+
+```
+════════════════════════════════════════════════════════════════
+  ⚠️ PR SIZE ALERT
+════════════════════════════════════════════════════════════════
+
+PR #23043 is getting large:
+
+  📊 Lines: 380 (Warning: > 200)
+  📁 Files: 12 (Critical: > 8)
+  👥 Teams: 3 (@YourOrg/team-a, @YourOrg/team-b, @YourCompany/core-oxide)
+  ⏱️ Est. Review: ~35 minutes
+
+════════════════════════════════════════════════════════════════
+  RECOMMENDATIONS
+════════════════════════════════════════════════════════════════
+
+1. 🔀 Split by CODEOWNERS (recommended)
+   Run: /split-pr 23043
+
+2. 📦 Split by feature
+   - PR-A: Display name support (core)
+   - PR-B: Tooltip updates (UI)
+   - PR-C: Analytics (tracking)
+
+3. 📝 Keep as-is with detailed description
+   Add clear sections explaining each change
+
+Choose: (1/2/3) or ignore
+```
+
+## 4. Process Change Detection
+
+When team processes change, I adapt.
+
+### What I Monitor
+
+-   PR template changes
+-   CODEOWNERS updates
+-   New lint rules
+-   Testing requirements
+-   Documentation requirements
+
+### Adaptation Output
+
+```
+════════════════════════════════════════════════════════════════
+  🔄 PROCESS CHANGE DETECTED
+════════════════════════════════════════════════════════════════
+
+I noticed changes in your project:
+
+1. **CODEOWNERS Updated**
+   - src/features/checkout/ now owned by @YourOrg/team-a-v2
+   - Will use new team for reviews
+
+2. **New Lint Rule Added**
+   - useRef must have explicit type
+   - Will check in self-review
+
+3. **PR Template Updated**
+   - Now requires "Breaking Changes" section
+   - Will include in PR creation
+
+Updating my workflow to reflect these changes.
+```
+
+## 5. Dependency Intelligence
+
+Detect when changes depend on other work.
+
+```
+════════════════════════════════════════════════════════════════
+  🔗 DEPENDENCIES DETECTED
+════════════════════════════════════════════════════════════════
+
+Your changes depend on:
+
+1. **Backend PR (BE-789)**
+   Status: ✅ Merged to staging
+   → Safe to proceed
+
+2. **Shared Types (@yourcompany-web/types)**
+   Status: ⚠️ PR #22900 pending
+   → May need to wait or use type override
+
+3. **Feature Flag (statsig)**
+   Status: ✅ Enabled for internal
+   → Can test on internal-com.yourcompany.io
+
+Recommendations:
+- Proceed with implementation
+- Test on staging (backend ready)
+- Note type dependency in PR description
+```
+
+## AI Execution
+
+These detections run automatically during `/full-flow`:
+
+1. **Before implementation**: Check AC for changes
+2. **During PR creation**: Detect size issues
+3. **When fetching ticket**: Identify multiple PRs
+4. **Throughout**: Monitor process changes
+
+I'll alert you when action is needed, otherwise proceed silently.

package/commands/learning/generate-memories.md

@@ -0,0 +1,270 @@
+---
+description: Convert learned patterns into persistent Cursor memories
+category: AI Self-Improvement
+aliases: [memories, remember, learn-patterns]
+---
+
+# Generate Memories - Persistent Learning
+
+Convert session learnings, user preferences, and established patterns into Cursor's memory system for cross-session persistence.
+
+## Usage
+
+```
+/generate-memories                    # Review and generate from current session
+/generate-memories --from-session     # Extract from this conversation
+/generate-memories --from-rules       # Extract key rules as memories
+/generate-memories --list             # Show existing memories
+/generate-memories --prune            # Remove outdated memories
+```
+
+## What Gets Memorized
+
+### User Preferences
+
+```
+"Always use useWatch instead of watch()"
+"Prefer AbortController for async cleanup"
+"Skip Confluence docs in full-flow"
+"Use Tampa Airport for testing"
+```
+
+### Project Patterns
+
+```
+"This project uses styled-components with S.* namespace"
+"Forms require OXForm wrapper"
+"Design tokens from core-oxide are mandatory"
+```
+
+### Common Decisions
+
+```
+"Empty string from API = treat as missing"
+"Use biggerThanSmall for mobile breakpoint"
+"Keep originalName for modal title preservation"
+```
+
+### Team Conventions
+
+```
+"PR size should be < 15 files"
+"Always add How to Test section"
+"Commit messages without ticket prefixes"
+```
+
+## Memory Categories
+
+| Category       | Examples                  | Persistence    |
+| -------------- | ------------------------- | -------------- |
+| **Preference** | "Always X over Y"         | Long-term      |
+| **Pattern**    | "Use this approach for Z" | Long-term      |
+| **Context**    | "Current project uses..." | Project-scoped |
+| **Temporary**  | "Currently working on..." | Session-only   |
+
+## Output Format
+
+```
+════════════════════════════════════════════════════════════════
+  MEMORY GENERATION
+════════════════════════════════════════════════════════════════
+
+📊 SESSION ANALYSIS
+
+Patterns detected in this session:
+  1. Used useWatch 5 times (consistent)
+  2. Applied AbortController pattern 3 times
+  3. Skipped Confluence docs
+  4. Used Tampa Airport for all test scenarios
+
+Preferences expressed:
+  1. "always use useWatch" (explicit)
+  2. "keep it simple" (implied)
+  3. "no ticket prefixes in commits" (correction)
+
+Corrections made:
+  1. Changed watch() → useWatch (I was wrong)
+  2. Removed hardcoded px values (reminded)
+
+════════════════════════════════════════════════════════════════
+  PROPOSED MEMORIES
+════════════════════════════════════════════════════════════════
+
+📝 NEW MEMORIES TO CREATE
+
+1. [Preference] React Hook Form
+   "Use useWatch hook instead of methods.watch() to prevent
+   unnecessary re-renders in form components."
+
+   Create? (y/n)
+
+2. [Pattern] Async Effect Cleanup
+   "All useEffect hooks with async operations must use
+   AbortController for cleanup and check signal.aborted
+   before setState."
+
+   Create? (y/n)
+
+3. [Preference] Test Locations
+   "Use Tampa Airport (TPA) as default location for testing
+   car rental flows."
+
+   Create? (y/n)
+
+4. [Preference] Commit Messages
+   "Keep commit messages natural and simple. No ticket
+   prefixes (fix:, feat:), no Jira IDs, no PR references."
+
+   Create? (y/n)
+
+════════════════════════════════════════════════════════════════
+  EXISTING MEMORIES TO UPDATE
+════════════════════════════════════════════════════════════════
+
+📝 MEMORY UPDATES
+
+1. [Update] "Common review issues in YourCompany codebase"
+   Current: Lists 5 issues
+   Add: "Empty string from API should be treated as missing"
+
+   Update? (y/n)
+
+════════════════════════════════════════════════════════════════
+  OUTDATED MEMORIES
+════════════════════════════════════════════════════════════════
+
+⚠️ POTENTIALLY OUTDATED
+
+1. "Use Redux for global state"
+   Last used: 6 months ago
+   Conflict: Project now uses React Query
+
+   Delete? (y/n)
+
+════════════════════════════════════════════════════════════════
+```
+
+## Memory Format
+
+Each memory should be:
+
+```markdown
+**Title:** {Short descriptive title}
+
+**Content:** {One paragraph, max 200 words}
+
+-   What the preference/pattern is
+-   When to apply it
+-   Brief rationale (optional)
+
+**Scope:**
+
+-   Global (all projects)
+-   Project (this workspace)
+-   Temporary (this session)
+```
+
+## AI Execution
+
+When user runs `/generate-memories`:
+
+### Step 1: Analyze Session
+
+```
+1. Review conversation history
+2. Identify patterns (repeated actions)
+3. Extract explicit preferences ("always...", "never...")
+4. Note corrections (I was wrong about...)
+```
+
+### Step 2: Categorize
+
+```
+1. User Preferences (explicit requests)
+2. Learned Patterns (consistent behavior)
+3. Corrections (mistakes to avoid)
+4. Context (project-specific)
+```
+
+### Step 3: Propose Memories
+
+```
+1. Draft memory for each finding
+2. Check against existing memories
+3. Identify updates vs new
+4. Identify conflicts
+```
+
+### Step 4: User Approval
+
+```
+1. Present each memory
+2. Get y/n approval
+3. Create/update via update_memory tool
+4. Confirm creation
+```
+
+## Integration with Self-Improve
+
+```
+┌─────────────────────────────────────────────────────────────┐
+│  SELF-IMPROVEMENT LOOP                                      │
+├─────────────────────────────────────────────────────────────┤
+│                                                             │
+│  Session Activity                                           │
+│       ↓                                                     │
+│  Pattern Detection (/self-improve)                          │
+│       ↓                                                     │
+│  Memory Generation (/generate-memories)                     │
+│       ↓                                                     │
+│  Cross-Session Persistence (Cursor Memories)                │
+│       ↓                                                     │
+│  Future Sessions (Auto-applied)                             │
+│                                                             │
+└─────────────────────────────────────────────────────────────┘
+```
+
+## Memory Lifecycle
+
+### Creation
+
+-   From explicit user request
+-   From repeated patterns
+-   From corrections
+
+### Usage
+
+-   Cited when applied: [[memory:ID]]
+-   Guides behavior silently
+-   Overridden by explicit user instruction
+
+### Maintenance
+
+-   Pruned if contradicted
+-   Updated if augmented
+-   Deleted if outdated
+
+## Best Practices
+
+1. **Quality over Quantity**
+    - Only memorize genuinely useful patterns
+    - Avoid obvious/trivial preferences
+2. **Be Specific**
+
+    - "Use useWatch for form fields" (good)
+    - "Forms are important" (too vague)
+
+3. **Include Rationale**
+
+    - "Use X because Y" is better than just "Use X"
+
+4. **Scope Appropriately**
+
+    - Global: True for all projects
+    - Project: Specific to this codebase
+    - Session: Just for current task
+
+5. **Review Periodically**
+    - `/generate-memories --prune` monthly
+    - Remove outdated patterns
+    - Update evolved practices