AgentSkillsCN

council-consensus

当用于Codex与Gemini的多视角分析,探索不同观点时使用。当出现“委员会视角”、“委员会的利弊”、“委员会评估选项”、“权衡是什么”、“委员会共识”时触发。

SKILL.md
--- frontmatter
name: council-consensus
description: Use for multi-perspective analysis with both Codex and Gemini exploring different viewpoints. Triggers on "council perspectives", "pros and cons from council", "council evaluate options", "what are the tradeoffs", "council consensus".

Council Consensus Skill

Multi-perspective analysis with both Codex and Gemini, each exploring for/against/neutral viewpoints for comprehensive evaluation.

When to Use

  • Evaluating technology choices
  • Architecture decisions
  • Trade-off analysis
  • When user needs pros and cons
  • Comparing different approaches
  • Decision-making support

Reasoning Level

high (default for balanced analysis)

Execution

  1. Identify the topic/question for consensus

  2. Gather relevant context

  3. For each model, request stance-based analysis:

    code
    Analyze this topic from multiple perspectives:
    
    Topic: <topic or decision>
    
    Please provide:
    1. **FOR** arguments - Best case for this approach
    2. **AGAINST** arguments - Concerns and risks
    3. **NEUTRAL** assessment - Objective analysis
    4. **Recommendation** - Your balanced conclusion
    
  4. Run BOTH commands in parallel:

    Codex:

    bash
    codex exec --sandbox read-only -c model_reasoning_effort="high" "<prompt>"
    

    Gemini:

    bash
    gemini -s -y -o json "<prompt>"
    
  5. Synthesize multi-model, multi-perspective analysis

Response Format

markdown
## AI Council Consensus Analysis

### Codex (GPT-5.2) Perspectives:

**FOR:**
- [Pro arguments]

**AGAINST:**
- [Con arguments]

**NEUTRAL:**
- [Objective observations]

**Recommendation:** [Codex's conclusion]

---

### Gemini Perspectives:

**FOR:**
- [Pro arguments]

**AGAINST:**
- [Con arguments]

**NEUTRAL:**
- [Objective observations]

**Recommendation:** [Gemini's conclusion]

---

### Council Consensus:

**Unanimous FOR:** [Points both models agree are advantages]

**Unanimous AGAINST:** [Concerns both models share]

**Divergent Views:** [Where the models disagree]

**Cross-Model Insights:**
- [Unique insights from combining perspectives]

**Council Recommendation:**
[Synthesized recommendation based on all perspectives from both models]

**Confidence Level:** [High if both agree, Medium if partial agreement, Low if divergent]

---
*Session IDs: Codex=[id], Gemini=[id]*