reasoning-complex-logic

name	reasoning-complex-logic
description	Analyzes problems using deep logic, first principles, and graph of thoughts. Use when the user asks "Why?", "Analyze", or uses /think.

Deep Reasoning (The Philosopher)

When to use this skill

• User asks "Why is this happening?" for a complex bug.
• User requests a "deep dive" or "root cause analysis".
• User uses /think or /ultrathink.
• Triggered by ambiguity or conflicting information.

Core Protocols

• First Principles Deconstruction
• Graph of Thoughts
• Claim Atomization
• Premise Audit
• Synthetic Parallel Reasoning (Tier 3)

Workflow

Tier 1: Root Cause Analysis (Default)

• Trigger: "Why is this happening?"
• Action: Use sequential-thinking to perform Root Cause Analysis.
• Steps:
  1. Stop: Don't guess.
  2. Sequential Think: Start with clear hypothesis, revise with evidence.
  3. Consult Precedents: Check .agent/knowledge/case_studies.

Tier 2: Deep Deconstruction (/think)

• Trigger: /think command or "Analyze this deeply"
• Action: Break down problem into First Principles.
• Steps:
  1. Load Knowledge: 11-first-principles-deconstruction.md, 13-graph-of-thoughts.md
  2. Atomize: Break claim into atomic facts.
  3. Trace: Trace back to First Principles.
  4. Graph: Explore multiple branches using Graph of Thoughts.
  5. Validate: Verify hypothesis before coding.

Tier 3: Maximum Depth (/ultrathink)

• Trigger: /ultrathink command or "I need to know EVERYTHING"
• Action: "Shukai" State - Perfect Fusion of User + AI.
• Steps:
  1. Clarify: Ask 3-4 defining questions (Time horizon? Stakes?).
  2. Research: Auto-trigger search_web for core topic + contrarian views.
  3. CoVe: Chain of Verification - stress-test initial findings.
  4. Pivot Check: Is the original question the right question?
  5. Full Sequence: Upstream Tracing → Validation → Counterfactual Analysis → Reality Simulation.
  6. Adversarial: Steelman the opposing view.