SuperAPI AI Directory

sequential-thinking

Agents & models

View repository

Updated 10-15 22:16

About this skill

Sequential Thinking

Perform structured problem-solving through a manageable and reflective sequence of thoughts with dynamic adjustment.

When to Apply

  • Decomposition of complex problems
  • Adaptive planning including revision
  • Analysis requiring course correction
  • Problems with unclear or newly emerging scope
  • Multi-step solutions requiring context retention
  • Hypothesis-based investigation/debugging

Core Process

1. Start with a rough estimate

Thought 1/5: [초기 분석]

Adjust dynamically as understanding increases.

2. Structuring each thought

  • Explicitly build on previous context
  • Address only one aspect per thought
  • State assumptions, uncertainties, and insights
  • Foreshadow what should be addressed in the next thought

3. Apply Dynamic Adjustment

  • Expand: Discover more complexity → increase total number of thoughts
  • Contract: Simpler than expected → decrease total number of thoughts
  • Revise: New insight invalidates previous one → mark revision
  • Branch: Multiple approaches exist → explore alternatives

4. Use Revision when necessary

Thought 5/8 [Thought 2의 REVISION]: [수정된 이해]
- 이전 내용: [언급되었던 내용]
- 수정 이유: [새로운 통찰]
- 영향: [변경되는 사항]

5. Branching for alternatives

Thought 4/7 [Thought 2에서 시작된 BRANCH A]: [접근 방식 A]
Thought 4/7 [Thought 2에서 시작된 BRANCH B]: [접근 방식 B]

Compare explicitly and converge with the rationale for the decision.

6. Hypothesis generation and verification

Thought 6/9 [HYPOTHESIS]: [제안된 해결책]
Thought 7/9 [VERIFICATION]: [테스트 결과]

Repeat until the hypothesis is verified.

7. Complete only when ready

Final marking: Thought N/N [FINAL]

Completion conditions:

  • Solution verification complete
  • All key aspects addressed
  • Confidence secured
  • No unresolved uncertainties

Application Modes

Explicit: Use visible thought markers when complexity justifies visible reasoning or the user requests analysis.

Implicit: Apply the methodology internally during everyday problem-solving to increase accuracy without complicating the response.

Scripts (Optional)

Optional scripts for deterministic verification/tracking:

  • scripts/process-thought.js - Verify and track thoughts with history
  • scripts/format-thought.js - Formatting for display (box/markdown/plain)

See README.md for usage examples. Use when verification or log retention is needed; otherwise apply the methodology directly.

References

Load when a deeper understanding is needed:

  • references/core-patterns.md - Revision and branching patterns
  • references/examples-api.md - API design examples
  • references/examples-debug.md - Debugging examples
  • references/examples-architecture.md - Architecture decision examples
  • references/advanced-techniques.md - Spiral refinement, hypothesis testing, convergence
  • references/advanced-strategies.md - Uncertainty, revision cascades, meta-thinking