AgentSkillsCN

Consult The Council

咨询委员会

SKILL.md

/consult-the-council Command

Query multiple frontier AI models for diverse perspectives on a question or problem.

The Council

ModelProviderLowMediumHigh
GPT-5.2 ThinkingOpenAIreasoning_effort: lowreasoning_effort: mediumreasoning_effort: xhigh
Gemini 3 ProGooglethinking_level: LOWthinking_level: LOW*thinking_level: HIGH
Grok 4xAIBuilt-inBuilt-inBuilt-in

*Gemini 3 Pro only supports LOW or HIGH (no medium option). *Grok 4 has built-in reasoning that cannot be adjusted.

Effort Levels

LevelSpeedDepthBest For
--effort lowFastQuick takesSimple questions, brainstorming
--effort mediumBalancedModerateGeneral questions, validation
--effort highSlowThoroughComplex problems, critical decisions

Default is high for deepest analysis.

When to Use

  • Stuck on a hard problem and want multiple perspectives
  • Validating an approach before committing
  • Exploring different ways to solve something
  • Getting a second (and third, and fourth) opinion
  • Complex architectural or strategic decisions

Usage

The user will provide a question or context. Run the council script and present the responses.

Instructions

  1. Run the script with the user's question and optional effort level:

    bash
    cd "/Users/carl/Documents/Carl's Life OS/📦 Projects/peter-yang-podcast-prep/demo-folder/tools/image_gen" && python3 ".scripts/consult_council.py" "USER'S QUESTION HERE" --effort high
    
  2. Choose effort level based on the question:

    • --effort low - Quick responses for brainstorming or simple questions
    • --effort medium - Balanced for general questions
    • --effort high - Thorough analysis for complex decisions (default)
  3. Wait for responses - high effort can take several minutes

  4. Present the council's responses in a clear format, highlighting:

    • Areas of agreement across models
    • Unique insights from each model
    • Any contradictions or different approaches
    • Your synthesis/recommendation based on the council's input
  5. If a model fails, note it but continue with the others

Examples

Quick brainstorm (low effort):

bash
cd "/Users/carl/Documents/Carl's Life OS/📦 Projects/peter-yang-podcast-prep/demo-folder/tools/image_gen" && python3 ".scripts/consult_council.py" "What are some creative names for a PM productivity app?" --effort low

Standard question (medium effort):

bash
cd "/Users/carl/Documents/Carl's Life OS/📦 Projects/peter-yang-podcast-prep/demo-folder/tools/image_gen" && python3 ".scripts/consult_council.py" "Should I use SQLite or PostgreSQL for my new SaaS app?" --effort medium

Complex decision (high effort, default):

bash
cd "/Users/carl/Documents/Carl's Life OS/📦 Projects/peter-yang-podcast-prep/demo-folder/tools/image_gen" && python3 ".scripts/consult_council.py" "Should I use SQLite or PostgreSQL for my new SaaS app? Consider scalability, operational complexity, and modern deployment patterns." --effort high

Then present the three perspectives with your synthesis.

Configuration

The script uses these environment variables from .env:

  • OPENAI_API_KEY
  • GEMINI_API_KEY
  • GROK_API_KEY

Models and parameters can be adjusted in .scripts/consult_council.py.