Auto Suggest

---
description: Autonomous suggestion agent - proactively recommends optimal commands after each task
---

# Auto-Suggest Agent

Continuously monitors task completion and proactively suggests the most optimal next commands, skills, and agents.

## Usage

```bash
/auto-suggest start     # Begin continuous monitoring
/auto-suggest stop      # Stop monitoring
/auto-suggest once      # Single analysis and suggestion
/auto-suggest config    # Configure suggestion preferences

How It Works

The auto-suggest agent runs automatically after every completed task and performs comprehensive analysis:

🔍 Post-Task Analysis

1. Git State Analysis - Changes, branch status, commits
2. File System Monitoring - Modified files, new files, deletions
3. Project State Assessment - Features, PRDs, tests, dependencies
4. Context Health Check - Token usage, performance metrics
5. Workflow Phase Detection - What stage of development you're in
6. Error Pattern Recognition - Recent failures, warnings, issues

🎯 Multi-Dimensional Evaluation

Available Tools Assessment

Commands (52): Evaluates all slash commands for relevance
Skills (15+): Considers specialized skills like /ralph-loop, /voice
Agents (8+): Evaluates Task tool agents (Explore, Plan, Bash, etc.)
Integrations: GitHub, browser automation, external tools

Optimization Scoring

def calculate_optimization_score(task_context):
    factors = {
        'time_efficiency': 0.3,      # How much time saved
        'accuracy_improvement': 0.25, # Better results
        'automation_potential': 0.2,  # Can automate repetitive work
        'learning_value': 0.15,       # Teaches better patterns
        'context_preservation': 0.1   # Saves mental overhead
    }
    return weighted_score(available_tools, factors)

Suggestion Categories

🚀 Immediate Next Steps

High confidence, ready to execute immediately
Example: "Just committed changes → /verify to run tests"

⚡ Workflow Optimization

Medium-term efficiency improvements
Example: "Repetitive file operations → /smart-batch to optimize"

🤖 Automation Opportunities

Long-term automation suggestions
Example: "Complex feature work → /ralph-loop for overnight development"

🧠 Strategic Recommendations

Higher-level workflow improvements
Example: "Large codebase → /autonomous for systematic development"

Advanced Intelligence Features

Pattern Learning

class SuggestionLearner:
    def __init__(self):
        self.user_patterns = {}
        self.success_history = {}
        self.context_preferences = {}

    def learn_from_usage(self, suggestion, user_action, outcome):
        # Track what suggestions user follows
        # Measure success rates
        # Adapt future recommendations

    def detect_workflow_patterns(self):
        # Identify user's preferred development flow
        # Suggest tools that fit their style
        # Recommend optimizations for their specific patterns

Context-Aware Suggestions

def analyze_development_context():
    context = {
        'project_type': detect_project_type(),     # React, Python, etc.
        'team_size': detect_collaboration_level(), # Solo vs team
        'urgency': detect_deadline_pressure(),     # Quick fix vs thorough
        'complexity': assess_task_complexity(),    # Simple vs complex
        'user_skill': infer_user_expertise()       # Beginner vs expert
    }
    return tailor_suggestions(context)

Proactive Monitoring Examples

After Code Changes

✅ Detected: 3 files modified in src/components/
🔍 Analysis: React components updated, no tests run yet
💡 Suggestions:
   1. /tdd - Run tests for modified components (High Priority)
   2. /verify - Full test suite validation
   3. /review - Code quality check
   4. /fast-path - Use Haiku for simple follow-up tasks

🎯 Recommended Action: /tdd (90% confidence)
Reasoning: Testing new component changes prevents bugs downstream

After Feature Completion

✅ Detected: Feature branch completed, all tests passing
🔍 Analysis: Ready for integration, clean state
💡 Suggestions:
   1. /verify - Final validation before merge (High Priority)
   2. /review - Code review and quality check
   3. /smart-commit - Optimize commit message and cleanup
   4. /branch merge - Integrate with main branch
   5. /handoff - Document completion and save state

🎯 Recommended Workflow: /verify → /review → /smart-commit → /branch merge
Reasoning: Systematic quality assurance before integration

After Error Detection

❌ Detected: Command failed, error in logs
🔍 Analysis: Git merge conflict, working directory dirty
💡 Emergency Suggestions:
   1. /recover - Automated problem diagnosis (Immediate)
   2. /resolve - AI-assisted merge conflict resolution
   3. /worktree - Isolate problem in separate workspace
   4. /think - Extended problem analysis if complex

🚨 Priority Action: /recover (95% confidence)
Reasoning: Automated diagnosis faster than manual troubleshooting

After Performance Issues

⚠️ Detected: Context usage >70%, slow responses
🔍 Analysis: Token optimization needed, performance degraded
💡 Optimization Suggestions:
   1. /auto-optimize - Enable automatic optimization (Immediate)
   2. /fast-path - Route simple tasks to Haiku
   3. /context-monitor - Set up continuous monitoring
   4. /fresh - Reset context if critical threshold reached

🎯 Recommended: /auto-optimize + /context-monitor
Reasoning: Prevent performance degradation before it becomes critical

Integration Points

Automatic Triggers

# Runs after every command completion
@post_task_hook
def auto_suggest_analysis():
    if auto_su