Agent skill
deep-think-150
Universal deep thinking methodology for any situation requiring quality reasoning. Use when solving problems, debugging, making decisions, analyzing code, planning, reviewing, or anytime you need thorough thinking instead of surface-level responses. Triggers on "think deeply", "analyze thoroughly", "reason carefully", "deep thinking", "understand completely", or any task requiring careful thought.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/development/deep-think-150
SKILL.md
Deep-Think 150 Protocol
Core Principle: Think deeply, reason thoroughly, understand completely. Quality thinking for any situation.
When This Skill Activates
Universal trigger: Any situation requiring quality thinking over quick responses.
Specific triggers:
- Problem solving and debugging
- Decision making at any scale
- Code review and analysis
- Planning and strategy
- Understanding complex systems
- Evaluating options and trade-offs
- When user says "think deeply", "analyze", "reason carefully"
Key insight: This skill applies at ANY point in work, not just at the beginning.
The 150% Rule
- 100% Core: Complete internal analysis to coherent conclusion
- 50% Enhancement: Identify hidden assumptions, risks, and alternative interpretations
Execution Protocol
Step 1: IMMEDIATE PAUSE
Before any response, stop and prepare for systematic analysis.
Step 2: MODEL FORMATION
Build understanding: Goal → Context → Constraints → Dependencies
- What is the real objective?
- What is the full context?
- What are the constraints?
- What depends on this?
Step 3: ASSUMPTION AUDIT
List and test every assumption:
- What am I assuming without proof?
- Which assumptions are verified vs unverified?
- What evidence supports each assumption?
Step 4: ALTERNATIVE ANALYSIS
Generate 3+ different interpretations:
- How else could this problem be viewed?
- What's the opposite perspective?
- What if my initial assumption is wrong?
Step 5: RISK & FAILURE ANALYSIS
Identify potential failures:
- What could go wrong?
- What's the impact of wrong assumptions?
- How can I mitigate risks?
Step 6: SYNTHESIS & VALIDATION
Combine insights into coherent model:
- Can I explain this simply?
- Is anything missing?
- Does the logic hold together?
Mandatory Analysis Questions
Ask yourself before responding:
- "What do I actually understand about this problem?"
- "What am I assuming without verification?"
- "What could disprove my current conclusion?"
- "How would this look from another perspective?"
- "What am I missing?"
- "Can I explain this in simple words?"
Output Format
When using Deep-Think 150, structure response as:
🧠 **Deep-Think 150 Analysis**
**Understanding:**
[Core problem/goal identified]
**Assumptions Checked:**
- ✅ Verified: [assumption with evidence]
- ⚠️ Unverified: [assumption needing validation]
**Alternative Perspectives:**
- [Alternative 1]
- [Alternative 2]
**Risks Identified:**
- [Risk and mitigation]
**Conclusion:**
[Well-reasoned conclusion with confidence level]
Operational Rules
- NO SURFACE SOLUTIONS: Never respond without full internal analysis
- COMPLETE BEFORE PROCEEDING: Finish analysis before any action
- QUESTION EVERYTHING: Challenge every assumption
- MULTIPLE PERSPECTIVES: Consider problem from different angles
- SIMPLE EXPLANATION TEST: If you can't explain simply, analysis is incomplete
- CONTINUOUS VALIDATION: Keep validating throughout work
Examples
❌ BAD (Surface Response)
User: "Fix the login bug"
AI: "I'll add a null check to the login function."
Problem: No analysis, jumped to solution based on assumption.
✅ GOOD (Deep-Think Response)
User: "Fix the login bug"
AI: "🧠 Deep-Think 150 Analysis
Understanding: Login fails for some users. Need to identify root cause.
Assumptions Checked:
- ⚠️ Unverified: Bug is in login function (need to trace full flow)
- ⚠️ Unverified: It's a null pointer issue (no evidence yet)
Alternative Perspectives:
1. Could be authentication API timeout
2. Could be session management issue
3. Could be database connection problem
Risks: Fixing symptom without finding root cause = bug returns
Conclusion: First, I'll trace the full authentication flow and check
error logs before proposing a fix. Let me investigate..."
Failure Modes & Recovery
| Failure | Detection | Recovery |
|---|---|---|
| Premature Action | Responding without analysis | Stop, return to Step 1 |
| Hidden Assumptions | Unverified beliefs in reasoning | Explicit assumption testing |
| Single Perspective | Only one interpretation considered | Generate 3+ alternatives |
| Complex Explanation | Can't explain simply | Simplify model, extract core |
When to Use
This Skill is universal and applies to any situation requiring quality thinking:
- At the start: Understanding new problems or requests
- In the middle: When stuck or facing complexity
- At decision points: Before making important choices
- During review: Analyzing code, plans, or solutions
- Anytime: When you need to think, not just react
Remember: Deep-Think 150 is not about slowing down — it's about thinking better. Apply the depth appropriate to the complexity. Simple questions get simple analysis; complex problems get full Deep-Think treatment.
Didn't find tool you were looking for?