AgentSkillsCN

31-analyze-think-150

【31】ANALYZE：无论何时，只要需要高质量的理性思考，都可运用这一通用的深度思维方法。适用于解决问题、调试代码、做出决策、分析代码、规划方案、审查项目，或在任何需要深入思考而非浅尝辄止的场合使用。当用户提出“深入思考”、“全面分析”、“谨慎推理”、“深度思考”、“彻底理解”，或在任何需要细致入微的思考任务中，此技能都将大显身手。

SKILL.md
--- frontmatter
name: 31-analyze-think-150
description: "[31] ANALYZE. 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."

Analyze-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:

  1. •"What do I actually understand about this problem?"
  2. •"What am I assuming without verification?"
  3. •"What could disprove my current conclusion?"
  4. •"How would this look from another perspective?"
  5. •"What am I missing?"
  6. •"Can I explain this in simple words?"

Output Format

When using Deep-Think 150, structure response as:

code
🧠 **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

  1. •NO SURFACE SOLUTIONS: Never respond without full internal analysis
  2. •COMPLETE BEFORE PROCEEDING: Finish analysis before any action
  3. •QUESTION EVERYTHING: Challenge every assumption
  4. •MULTIPLE PERSPECTIVES: Consider problem from different angles
  5. •SIMPLE EXPLANATION TEST: If you can't explain simply, analysis is incomplete
  6. •CONTINUOUS VALIDATION: Keep validating throughout work

Examples

❌ BAD (Surface Response)

code
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)

code
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

FailureDetectionRecovery
Premature ActionResponding without analysisStop, return to Step 1
Hidden AssumptionsUnverified beliefs in reasoningExplicit assumption testing
Single PerspectiveOnly one interpretation consideredGenerate 3+ alternatives
Complex ExplanationCan't explain simplySimplify 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

Session Log Entry (MANDATORY)

After completing this skill, write to .sessions/SESSION_[date]-[name].md:

code
### [date - HH"MM] Analyze-Think 150 Complete
**Topic:** <what was analyzed>
**Conclusion:** <key insight/decision>
**Risks:** <identified risks>

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.