---

name: learn description: Provide feedback on CYNIC judgments to improve future evaluations. Use when asked to correct a judgment, provide feedback, mark as correct/incorrect, or help CYNIC learn from mistakes. user-invocable: true

/learn - CYNIC Feedback Loop

"φ learns from φ"

Quick Start

Copy/learn <judgment_id> correct|incorrect [reason]

What It Does

Provides feedback to improve CYNIC's judgment accuracy:

• Mark judgments as correct/incorrect/partial
• Explain why the judgment was wrong
• Suggest what the score should have been
• Trigger weight calibration

Feedback Types

Examples

Mark Correct

Copy/learn jdg_abc123 correct

Mark Incorrect with Reason

Copy/learn jdg_abc123 incorrect "missed security vulnerability"

Provide Score Correction

Copy/learn jdg_abc123 incorrect --score 35 "should have been BARK not WAG"

Implementation

Use the brain_cynic_feedback MCP tool:

Copybrain_cynic_feedback({
  judgmentId: "jdg_abc123",
  outcome: "correct|incorrect|partial",
  reason: "explanation of feedback",
  actualScore: 35  // What it should have been (0-100)
})

Learning System

CYNIC uses feedback to:

1. Calibrate Weights: Adjust dimension weights
2. Detect Biases: Find systematic errors
3. Improve Accuracy: Learn from mistakes

Check learning state:

Copybrain_learning({ action: "state" })

View detected biases:

Copybrain_learning({ action: "biases" })

Trigger calibration:

Copybrain_learning({ action: "calibrate" })

Impact

See Also

• /judge - Create judgments to learn from
• /patterns - See learned patterns
• /health - Check learning system status