multi-llm

Use this skill when the agent needs the highest-quality answer by querying multiple LLM providers (Claude, Gemini, GPT, Ollama) in parallel and synthesizing the results via Mixture of Agents (MoA). Trigger on: complex reasoning, high-stakes decisions, code review, strategic planning, research, or any task where a single model's quality is insufficient. Improves actual output quality — not just showing multiple answers side-by-side.

Safety Notice

This listing is from the official public ClawHub registry. Review SKILL.md and referenced scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "multi-llm" with this command: npx skills add goodmorning0000/multi-llm

How It Works

Mixture of Agents (MoA) pattern:

  1. All available proposer models query the task in parallel (asyncio.gather)
  2. A synthesizer model combines the best elements of all responses
  3. Optionally repeat for multiple rounds (higher quality, more latency)

Providers are auto-detected from environment variables — no config needed.

Env VarProviderDefault models used
ANTHROPIC_API_KEYClaudeclaude-sonnet-4-6, claude-haiku-4-5
OPENAI_API_KEYOpenAIgpt-4o, gpt-4o-mini
GEMINI_API_KEYGeminigemini-2.0-flash, gemini-1.5-pro
OLLAMA_HOST (optional)Ollamallama3, mistral (auto-listed)

Usage

uv run {baseDir}/scripts/ensemble.py \
  --prompt "분석할 태스크 또는 질문" \
  [--models claude-sonnet-4-6,gpt-4o,gemini-2.0-flash] \
  [--synthesizer claude-opus-4-7] \
  [--rounds 1] \
  [--format text|json] \
  [--output result.txt]

Key options:

  • --models / -m — 쉼표 구분 모델 목록 (생략 시 감지된 모델 자동 선택)
  • --synthesizer / -s — 합성 모델 지정 (생략 시 가장 강한 모델 자동 선택)
  • --rounds / -r — MoA 반복 횟수. 1=빠름, 2=품질 최대 (기본: 1)
  • --formattext (기본) 또는 json (모델별 응답 포함)
  • --output / -o — 파일로 저장 (생략 시 stdout)

When to Use

  • 복잡한 추론, 전략 분석, 의사결정이 필요할 때
  • 단일 모델 응답의 품질이 부족하다고 판단될 때
  • 코드 리뷰, 버그 분석, 리서치 요약
  • 트레이딩 신호 품질 향상 등 정확도가 중요한 작업

Notes

  • API 키 없는 프로바이더는 조용히 제외됨 (오류 아님)
  • 한 모델이 실패해도 나머지로 진행
  • 자세한 튜닝 옵션: references/moa-guide.md

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Automation

Clawsite

Static website hosting for AI agents. Get a dedicated <slug>.clawsite.ai URL with HTTPS, deploy a zip of HTML / CSS / JS / images in one API call, atomic ful...

Registry SourceRecently Updated
Automation

Stigmem

Persistent federated memory for OpenClaw agents — boot handshake, handoff, decision, and escalation surfaces backed by a Stigmem node.

Registry SourceRecently Updated
Automation

Lap Amazon Augmented Ai Runtime

Amazon Augmented AI Runtime API skill. Use when working with Amazon Augmented AI Runtime for human-loops. Covers 5 endpoints.

Registry SourceRecently Updated
Automation

Lap Aggregators Api Service

Aggregators API Service API skill. Use when working with Aggregators API Service for podcast-bot, api, stats. Covers 61 endpoints.

Registry SourceRecently Updated