sequential-thinking
Sequential Thinking
Use When: Complex problems requiring structured analysis
Core Pattern
thought_structure[5]{step,format,purpose}:
1,Initial analysis,Understand the problem space
2,Build on previous,Deepen understanding
3,[REVISION],Correct earlier assumptions
4,[BRANCH],Explore alternatives
5,[FINAL],Synthesize solution
Output Format
Thought 1/N: [Initial analysis of the problem]
- Key observations
- Initial assumptions
Thought 2/N: [Building on Thought 1]
- Deeper analysis
- Connections discovered
Thought 3/N [REVISION of Thought 1]: [Corrected understanding]
- What was wrong
- Corrected view
Thought 4/N [BRANCH A]: [Alternative approach]
- Different angle
- Trade-offs
Thought 5/N [FINAL]: [Solution synthesis]
- Recommended approach
- Key decisions
Dynamic Adjustment
adjustment[3]{action,when,how}:
Expand,Complexity increases,Add more thoughts (N+1)
Contract,Simpler than expected,Skip to FINAL
Branch,Multiple valid paths,Create BRANCH A/B/C
When to Use
use_cases[5]{scenario,thoughts,model}:
Architecture design,5-8,opus
Bug root cause,3-5,sonnet
Performance optimization,4-6,sonnet
Security analysis,5-7,opus
Refactoring strategy,4-6,sonnet
Integration
phase_integration[4]{phase,use,focus}:
Phase 1,Requirements analysis,Understand scope and constraints
Phase 2,Technical planning,Evaluate approaches
Phase 3,Complex implementation,Debug tricky issues
Phase 4,Code review,Deep analysis of changes
Example
User: "The login is slow but I don't know why"
Thought 1/4: [Initial analysis]
- Login involves: UI → API → Auth → DB
- "Slow" is vague - need metrics
- Possible bottlenecks: network, DB, auth provider
Thought 2/4: [Building on observations]
- Auth provider (external) is most likely
- DB queries should be fast (indexed)
- Network latency is user-dependent
Thought 3/4 [REVISION of Thought 2]: [After checking logs]
- Actually, DB query is taking 2.3s
- Missing index on users.email
- Auth provider is fine (50ms)
Thought 4/4 [FINAL]: [Solution]
- Add index: CREATE INDEX idx_users_email ON users(email)
- Expected improvement: 2.3s → 50ms
- Verify with EXPLAIN ANALYZE
Invoke: Use for problems requiring structured exploration before solution.
More from nguyenthienthanh/aura-frog
stitch-design
Generate UI designs using Google Stitch AI with optimized prompts
36angular-expert
Angular 17+ gotchas and decision criteria. Covers signals vs observables, standalone patterns, and common pitfalls Claude gets wrong.
31flutter-expert
Flutter/Dart mobile expert. PROACTIVELY use when working with Flutter, Dart, mobile apps. Triggers: flutter, dart, widget, bloc, riverpod
16seo-expert
SEO and search engine optimization expert. PROACTIVELY use when working with meta tags, structured data, Core Web Vitals, sitemap, robots.txt, canonical URLs. Triggers: SEO, meta tags, schema markup, search ranking
13dev-expert
Development patterns for React, Vue, Laravel, Next.js, React Native - state management, forms, API integration
13react-native-expert
React Native best practices expert. PROACTIVELY use when working with React Native, mobile apps, Expo. Triggers: react-native, expo, mobile, iOS, Android, NativeWind
13