DeepThink

Adaptive deep reasoning skill — auto-detects question complexity and adjusts analysis depth. Quick mode for simple queries, full protocol (ToT / Socratic / verification / iteration) for complex problems. Bilingual (auto Chinese/English).

Install

npx skills add wxhou/deepthink

Works with Claude Code, Cursor, Codex, GitHub Copilot, Windsurf, Gemini, Cline, Roo, Zed, VS Code and 20+ other agents listed at skills.sh.

Usage

# Claude Code
/deepthink 选择 PostgreSQL 还是 MongoDB?

# Cursor / Codex / Copilot / others
deepthink: 选择 PostgreSQL 还是 MongoDB?

How It Works

Quick Mode (default for simple queries)

If the question is answerable in one sentence → 1-3 sentence direct answer, stop.

Deep Mode (default for complex queries)

1. Problem Decomposition + Assumption Statement — mandatory list of key assumptions with self-verification (✅ holds / ❌ fails / ❓ pending)
2. Effort Level — low (2-3 rounds) / medium (5-6) / high (7-9), adjusted by complexity rules
3. Socratic Questioning — Clarify / Assumptions / Evidence / Counterexamples / Alternatives / Consequences
4. Multi-Level Analysis — Tree of Thoughts (mandatory for high effort) → Understand → Plan → Execute → Verify
5. Verification — Reflection / First Principles / Backward / Self-Consistency
6. Iteration — if Low/Medium confidence or multi-subsystem
7. Completeness Check — confirm all steps before output

Output Format

---
## 🤔 DeepThink Analysis

### Core Problem
[1-sentence summary of the real question]

### Key Assumptions
[List with ✅/❌/❓]

### Conclusion
[Final answer]

### One-Sentence Summary
[Core recommendation]

### Confidence: [High/Medium/Low]
---

Repository Layout

deepthink/
├── SKILL.md             # Skill entry (YAML frontmatter + usage)
├── core/                # Canonical protocol source (platform-agnostic)
│   ├── protocol.md      # English
│   └── protocol.zh.md   # Chinese
├── references/          # Same content as core/, kept for Claude Code compat
│   ├── en.md
│   └── zh.md
├── evals/
│   └── evals.json       # 12 test cases (quick/medium/high)
└── scripts/
    └── run_evals.py     # Eval runner

Evals

12 test cases across quick / decision / debug / tradeoff / arch / concept / complex types:

python scripts/run_evals.py --iter 1

Outputs go to deepthink-workspace/iteration-N/eval-{id}-{type}/with_skill/outputs/output.txt.

License

MIT

Version

5.1.0 — published to skills.sh