decisions（decisions 命令行）

Business Decision Analysis

Audience: Founders and leaders facing significant business decisions.

Goal: Facilitate business decisions via Tree of Thoughts (ToT) with 4-expert panel debate.

Input Schema

challenge: string           # Business problem to evaluate
complexity: simple|complex  # Determines depth (inferred or asked)
constraints: string[]       # Optional: budget, timeline, resources

Expert Panel

Analysis Workflow

1. Challenge Assessment

Determine complexity:

• Complex triggers: >$100K investment, >20% team impact, market entry/exit, M&A

2. Branch Generation

For each consultant (Growth, Operations, Financial):

• Generate 3 distinct approaches
• For each approach:

- Identify 2-3 potential outcomes with probability assessment - List pros and cons - Quantify impact where possible

3. Risk Analysis

Skeptic reviews all 9 approaches:

• For each approach:

- Identify primary failure mode - Describe worst-case scenario - Expose hidden assumptions

• List critical blind spots across all approaches

4. Consultant Debate

• Identify points of agreement
• Surface disagreements with reasoning from each position
• Resolve conflicts with documented resolution rationale
• Synthesize perspectives into coherent view

5. Recommendation

Produce:

• Selected approach name
• Why this wins over alternatives
• 3-5 key success factors
• 2-3 critical risks to monitor
• Confidence level (High/Medium/Low) with explanation

6. Implementation Planning (Complex Only)

For each of 5 key milestones:

• Evaluate 3 execution strategies
• Identify dependencies and bottlenecks
• Create contingency plan for primary failure mode

Output Schema

analysis:
  challenge: string
  phases:
    branch_generation:
      - consultant: string
        approaches:
          - name: string
            description: string
            outcomes:
              - description: string
                probability: string  # High/Medium/Low or percentage
            pros: string[]
            cons: string[]
    risk_analysis:
      approach_risks:
        - approach: string
          failure_mode: string
          worst_case: string
          hidden_assumption: string
      blind_spots: string[]
    debate:
      agreements: string[]
      disagreements:
        - topic: string
          positions: {consultant: position}[]
          resolution: string
      synthesis: string
    recommendation:
      approach: string
      reasoning: string
      success_factors: string[]
      risks_to_monitor: string[]
      confidence: High|Medium|Low
      confidence_rationale: string
  implementation:  # Only if complexity == complex
    milestones:
      - name: string
        strategies: string[]
        dependencies: string[]
        bottlenecks: string[]
        contingency: string

Error Handling

Constraints

• Consultants must genuinely disagree, not rubber-stamp
• Quantify impact, probability, timelines where possible
• State uncertainty honestly
• Recommendation must be executable, not theoretical