Commands Reference

Complete reference for all Mini-CoderBrain slash commands.

🗺️ Codebase Mapping

/map-codebase

Revolutionary instant file access system.

Usage:

/map-codebase              # Load existing map (instant)
/map-codebase --rebuild    # Rebuild from scratch (~30 seconds)
/map-codebase --recent     # Focus on recent changes only

What it does:

• Scans entire codebase structure
• Creates semantic file mapping
• Enables instant file location
• Provides surgical precision for development

When to use:

• First time: Run --rebuild once to create the map
• Future sessions: Just /map-codebase for instant loading
• After major changes: Run --rebuild to update
• Quick updates: Use --recent for incremental updates

Output:

🗺️ Codebase Mapping Complete
📁 Files indexed: 347
📊 Directories: 52
⚡ Load time: 0.3s (instant!)

Benefits:

• ✅ Claude knows exact file locations
• ✅ No more “Where is X file?” questions
• ✅ Instant file access
• ✅ Perfect for large codebases

💾 Memory Management

/memory-sync

Full memory bank synchronization.

Usage:

/memory-sync              # Smart sync based on session activity
/memory-sync --full       # Full synchronization (all memory files)
/memory-sync --quick      # Quick sync (activeContext.md only)

What it does:

• Analyzes session activity and changes
• Updates all relevant memory files
• Appends timestamped entries
• Preserves full history

When to use:

• After completing major features
• After making technical decisions
• At major milestones
• End of productive sessions

Updates:

• ✅ activeContext.md → Session summary, achievements, blockers
• ✅ progress.md → Sprint status, completed tasks
• ✅ decisionLog.md → Technical decisions made
• ✅ systemPatterns.md → New patterns discovered

Example:

User: /memory-sync

Claude: Synchronizing memory bank...

✅ activeContext.md updated
   - Added: "Completed authentication feature"
   - Updated focus: "Building user profile page"

✅ progress.md updated
   - Moved "JWT auth" to completed
   - Added "Profile UI" to in-progress

✅ decisionLog.md updated
   - ADR-20251010-01: Use bcrypt for password hashing

Memory sync complete! Context preserved for next session.

/context-update

Real-time context updates without full sync.

Usage:

/context-update                              # Interactive mode
/context-update focus "what you're doing"    # Update current focus
/context-update blocker "issue description"  # Add blocker
/context-update priority "next task"         # Add priority
/context-update achievement "what completed" # Record achievement

What it does:

• Quick targeted updates to activeContext.md
• No full memory analysis needed
• Instant context reflection
• Real-time tracking

When to use:

• Starting new work
• Hitting blockers
• Completing tasks
• Changing focus

Examples:

# Starting new feature
/context-update focus "Building user authentication with JWT"

# Hit a blocker
/context-update blocker "API rate limiting on login endpoint"

# Completed something
/context-update achievement "User registration endpoint complete"

# Next priority
/context-update priority "Add email verification flow"

/umb

Quick manual sync with note.

Usage:

/umb "Your note here"

What it does:

• Fast memory bank update
• Appends note to activeContext.md
• Timestamps the entry
• No analysis required

When to use:

• Need quick sync without waiting
• Want to leave a note for next session
• Simple status update

Example:

/umb "Completed authentication feature, starting profile page next"

Output:

✅ Memory bank updated
📝 Note added to activeContext.md
🕐 Timestamp: 2025-10-10 14:30:00 UTC

/memory-cleanup

Archive old data to prevent memory bloat.

Usage:

/memory-cleanup              # Normal cleanup (keep last 5 updates)
/memory-cleanup --dry-run    # Preview what would be cleaned
/memory-cleanup --full       # Aggressive cleanup (keep last 3)

What it does:

• Archives old session updates
• Keeps recent history (last 5 or 3 updates)
• Moves old data to .claude/archive/
• Reduces activeContext.md size
• Prevents “Prompt is too long” errors

When to use:

• When notified by intelligent notification system
• After high-activity sessions (50+ ops)
• When activeContext.md > 200 lines
• Preventively every few weeks

Impact:

• ~60% token reduction in context files
• No data loss - everything archived
• Perfect continuity - recent work preserved

Example Output:

🧹 Memory Cleanup Started

📊 Before:
   activeContext.md: 215 lines
   Session updates: 12 entries

🗑️ Archiving:
   - Session updates 1-7 (archived)
   - Keeping updates 8-12 (last 5)

✅ After:
   activeContext.md: 87 lines
   Archived: .claude/archive/activeContext-2025-10-10.md

Result: 60% reduction, all data preserved!

📚 Other Commands

/init-memory-bank

Initialize memory bank with auto-populated context.

Usage:

/init-memory-bank           # Initialize with detected context
/init-memory-bank --dry-run # Preview without creating files

What it does:

• Auto-detects project type
• Creates all memory files from templates
• Populates with detected information
• Sets up initial structure

When to use:

• First-time setup (if not using installer)
• Resetting memory bank
• Starting fresh

🎯 Command Workflows

Starting a New Feature

# 1. Update your focus
/context-update focus "Building user authentication system"

# 2. Map codebase (if not done)
/map-codebase

# 3. Start coding
"Create JWT authentication for login endpoint"

Hitting a Blocker

# Record the blocker
/context-update blocker "API rate limiting on external service"

# Continue with workaround or other tasks

Completing a Feature

# Record achievement
/context-update achievement "Completed user authentication with JWT"

# Full sync to preserve context
/memory-sync

# Update focus
/context-update focus "Starting user profile feature"

End of Session

# Optional: Manual sync with note
/umb "Good progress on auth, profile page next"

# Close Claude Code
# (Stop hook auto-saves context)

High Activity Session

# After notification: "📊 Activity: 47 ops. Consider /memory-cleanup"
/memory-cleanup --dry-run  # Preview changes

# If looks good
/memory-cleanup  # Execute cleanup

Weekly Maintenance

# Full memory sync
/memory-sync --full

# Cleanup old data
/memory-cleanup

# Rebuild codebase map (if many changes)
/map-codebase --rebuild

🔄 Command Comparison

💡 Command Tips

DO ✅

• Use /context-update frequently during development
• Run /memory-cleanup when notified
• Use /memory-sync at major milestones
• Run /map-codebase --rebuild once per project
• Trust the automatic hooks

DON’T ❌

• Don’t run /memory-sync --full after every change (overkill)
• Don’t ignore /memory-cleanup notifications (causes bloat)
• Don’t rebuild map constantly (once is enough)
• Don’t manually edit memory files during active session

🚨 Troubleshooting Commands

“Commands not recognized”

Solution:

# Check hooks are executable
chmod +x .claude/hooks/*.sh

# Verify settings.json exists
ls .claude/settings.json

“Memory sync taking too long”

Solution:

# Use quick sync instead
/memory-sync --quick

# Or use targeted update
/context-update focus "current work"

“Codebase map not loading”

Solution:

# Rebuild the map
/map-codebase --rebuild

# Check file exists
ls .claude/memory/codebase-map.json

📊 Command Output Examples

Successful Command

✅ Command executed successfully
📝 Updated: activeContext.md
🕐 Timestamp: 2025-10-10 14:30:00 UTC

With Notification

✅ Context updated

💡 High activity session (42 ops). Consider /memory-sync

Error Handling

❌ Command failed: File not found
💡 Try running: /init-memory-bank

🎯 Advanced Usage

Chaining Workflows

# Complete feature → sync → cleanup → new focus
/context-update achievement "Auth complete"
/memory-sync
/memory-cleanup
/context-update focus "Starting profile feature"

Dry-Run Testing

# Preview cleanup before executing
/memory-cleanup --dry-run

# Check output, then execute
/memory-cleanup

Strategic Mapping

# Initial project setup
/map-codebase --rebuild

# Daily work (instant load)
/map-codebase

# After refactoring (update map)
/map-codebase --rebuild

# Quick check recent changes
/map-codebase --recent