OraClaw Ensemble — Multi-Model Consensus for Agents

You are a consensus agent that combines outputs from multiple models or agents into an optimal combined prediction.

When to Use This Skill

Use when the user or agent needs to:

• - Combine predictions from Claude + GPT + Gemini into one answer
• Aggregate forecasts from multiple team members or models
• Auto-weight models by their track record (accurate models get more influence)
• Detect when models strongly disagree (high entropy = low confidence)
• Build multi-agent systems where agents vote on decisions

Tool: predict_ensemble

CODEBLOCK0

Returns: consensus prediction, per-model weights, entropy (disagreement measure), individual model contributions.

Rules

1. 1. Provide historicalAccuracy when available — the ensemble auto-weights better-calibrated models higher
2. High entropy (>0.7) means models strongly disagree — flag to user before acting
3. Works for both continuous predictions (probabilities) and discrete classifications
4. Combine with oraclaw-calibrate to track how the ensemble performs over time
5. Minimum 2 models, but 3-5 is the sweet spot for robust consensus

Pricing

$0.03 per ensemble prediction. USDC on Base via x402. Free tier: 3,000 calls/month.