ThinkDeep: Extended Reasoning Analysis
Step 1: Parse Request Parameters
Analyze the thinkdeep request: $ARGUMENTS
Parameter Extraction:
-
Main prompt: [extract core thinking request]
-
Focus areas: [extract "architecture, performance, security" etc]
-
File references: [extract "reference to file.py" patterns]
-
Problem context: [extract additional context provided]
Parsed Configuration:
-
Prompt: [core request]
-
Focus areas: [identified areas]
-
File references: [any files mentioned]
-
Problem context: [additional context]
Step 2: Deep Sequential Analysis
Use sequential thinking (mcp__sequential-thinking__sequentialthinking) for extended reasoning:
Thought 1: Context Understanding
Analyze the current session context to understand:
-
What has been discussed so far
-
Previous decisions and analyses made
-
User's specific constraints and goals
-
Existing solutions or approaches tried
Thought 2: Problem Decomposition
Break down the core issue into:
-
Primary problem statement
-
Key sub-problems or components
-
Underlying assumptions and constraints
-
Success criteria and desired outcomes
Thought 3: Assumption Challenge
Question the underlying assumptions:
-
What assumptions are we making?
-
Are these assumptions valid in this context?
-
What if the opposite were true?
-
Alternative perspectives to consider
Thought 4: Alternative Approaches Exploration
Explore different solution paths:
-
Conventional approaches and their pros/cons
-
Unconventional or innovative solutions
-
Trade-offs between different options
-
Hybrid or combined approaches
Thought 5: Context Integration
Incorporate relevant context:
-
File references and code patterns
-
Previous session insights
-
User's specific technical stack
-
Business or project constraints
Thought 6: Edge Case Analysis
Identify potential issues:
-
What could go wrong with each approach?
-
Edge cases and failure modes
-
Scalability and performance concerns
-
Security and maintenance implications
Thought 7: Risk Assessment
Evaluate risks for each option:
-
Implementation complexity
-
Technical debt potential
-
Team skill requirements
-
Timeline and resource impacts
Thought 8: Solution Synthesis
Combine insights into recommendations:
-
Primary recommended approach
-
Backup or alternative options
-
Implementation considerations
-
Next steps and validation points
Continue sequential thinking until satisfaction, using:
-
isRevision: true to refine previous thoughts
-
branchFromThought to explore alternative reasoning paths
-
needsMoreThoughts: true to extend analysis if needed
Step 3: Critical Evaluation & Synthesis
Self-Critique Questions:
-
Does the analysis address the user's specific context?
-
Are the recommendations practical and implementable?
-
Have we considered the most important constraints?
-
Are there any blind spots or missing perspectives?
Contextual Validation:
-
Fit with user's technical stack and preferences
-
Alignment with previous decisions in session
-
Practicality given team skills and resources
-
Integration with existing systems or patterns
Final Synthesis:
-
Primary Recommendation: [best approach with rationale]
-
Alternative Options: [backup plans with trade-offs]
-
Implementation Guidance: [key considerations and next steps]
-
Risk Mitigation: [how to address identified concerns]
Usage Examples
Basic deep analysis:
/atelier-thinkdeep "Analyze my authentication architecture"
With file references:
/atelier-thinkdeep "Evaluate my database schema with reference to models/user.py"
With focus areas:
/atelier-thinkdeep "Assess my microservices design focusing on performance and security"
Complex architectural decisions:
/atelier-thinkdeep "Should I use GraphQL or REST for my API considering team skills and scalability needs"
When to Use ThinkDeep
-
Architecture decisions: When choosing between major design approaches
-
Complex problem solving: Multi-faceted issues requiring deep analysis
-
Assumption validation: Challenging conventional wisdom or previous decisions
-
Alternative exploration: When stuck and need fresh perspectives
-
Risk assessment: Before major implementation decisions
ThinkDeep vs Challenge
Use /atelier-thinkdeep: Deep exploration, comprehensive analysis, alternative discovery, complex decisions Use /atelier-challenge: Question assumptions, test validity, assess risks, prevent automatic agreement
Key distinction: ThinkDeep = deep exploration, Challenge = critical evaluation
Expected Outcomes
After using this command:
-
Comprehensive understanding of the problem space
-
Multiple approaches evaluated with trade-offs
-
Clear recommendation with solid reasoning
-
Risk assessment and mitigation strategies
-
Actionable next steps for implementation