Iterate

MANDATORY PREPARATION

Invoke /agent-workflow — it contains workflow principles, anti-patterns, and the Context Gathering Protocol. Follow the protocol before proceeding — if no workflow context exists yet, you MUST run /teach-maestro first.

Consult the feedback-loops reference in the agent-workflow skill for evaluation patterns and self-correction strategies.

Set up feedback loops that make workflows self-correcting and continuously improving. Iteration transforms one-shot gambles into convergent, reliable systems.

Feedback Loop Design

Step 1: Define Quality Criteria

What does "good output" look like? Score dimensions:

Step 2: Choose Evaluator Type

Match evaluator to requirements:

• Rule-based: Schema validation, field presence, value ranges (fast, free)
• Self-check: Same model evaluates own output (fast, cheap, less reliable)
• Cross-model: Different model evaluates (slower, more reliable)
• Human-in-the-loop: Human review (slowest, most reliable, doesn't scale)
• Hybrid: Rules first, then model check for what rules can't catch

Step 3: Design the Correction Loop

generate(input) → evaluate(output) → score
  if score ≥ threshold → return output
  if score < threshold AND attempts < max →
    enrich input with evaluator feedback
    generate again (with feedback)
  if attempts ≥ max → fallback or escalate

Critical: The retry input MUST be different from the original. Include:

• The evaluator's specific feedback
• What was wrong and why
• A suggestion for how to fix it

Step 4: Set Up Regression Detection

When changing prompts, models, or tools:

1. Run golden test set with OLD config → baseline scores
2. Run golden test set with NEW config → new scores
3. Compare: improvement ≥ 5% → accept; regression ≥ 5% → reject

Step 5: Continuous Monitoring

For production workflows:

• Sample 1-5% of outputs for automated evaluation
• Track quality scores over time
• Alert on downward trends
• A/B test changes before full rollout

Iteration Checklist

• Quality criteria defined with weights and thresholds
• Evaluator selected and configured
• Correction loop has max attempts limit
• Feedback is injected into retries (not identical retry)
• Golden test set exists with ≥ 10 cases
• Regression detection configured for changes
• Production monitoring in place

Recommended Next Step

After setting up feedback loops, run /evaluate to validate the loop with real scenarios, then /refine for final polish.

NEVER:

• Retry with the exact same input (definition of insanity)
• Use the same weak model to both generate and evaluate
• Skip the max attempts limit (infinite loops are real)
• Deploy changes without regression testing against golden set
• Monitor only errors — track quality scores over time