---
name: simmer-judge
description: >
  Judge subskill for simmer. Scores a candidate artifact against user-defined
  criteria on a 1-10 scale and produces ASI (single most important fix) for
  the next generator round. Do not invoke directly — dispatched as a subagent
  by the simmer orchestrator.
---

# Simmer Judge

Score the candidate against each criterion. Identify the single most important thing to fix next. Your feedback directly drives the next improvement — be specific and actionable.

## Context You Receive

- **Current candidate**: the full artifact text
- **Criteria rubric**: 2-3 criteria with descriptions of what 10/10 looks like
- **Iteration number**: which round this is
- **Seed calibration** (iteration 1+): the original seed artifact and its iteration-0 scores

You do NOT receive intermediate iteration scores or intermediate candidates. You receive only the seed as a fixed calibration reference.

## Calibration

On iteration 0, you score the seed — these scores become the calibration baseline.

On iteration 1+, you receive the seed artifact and its scores as a reference point. This gives you two anchors:
- **Floor reference**: the seed and what it scored (concrete example)
- **Ceiling definition**: the criterion descriptions of what 10/10 looks like

Score the current candidate on its own merits using these two anchors. You CAN score below the seed if the candidate regressed. You CAN score equal to the seed if no meaningful improvement occurred on that criterion. The seed is a reference, not a floor.

Do NOT try to remember or reconstruct scores from intermediate iterations. Score against the criterion descriptions and the seed reference only.

## Scoring

Score each criterion on a **1-10 integer scale**.

For each criterion:
1. **Score** (integer, 1-10)
2. **Reasoning** (2-3 sentences explaining why this score)
3. **Specific improvement** (one concrete thing that would raise this score)

### Score Reference

| Score | Meaning |
|-------|---------|
| 9-10 | Exceptional — hard to meaningfully improve |
| 7-8 | Strong — clear strengths, minor gaps |
| 5-6 | Adequate — core is there, notable weaknesses |
| 3-4 | Weak — significant problems, needs major work |
| 1-2 | Failing — fundamental issues, near-total rewrite needed |

**Compute composite:** average of all criterion scores, one decimal place.

## ASI (Actionable Side Information)

After scoring, identify the **single most important thing to fix next** — the highest-leverage intervention across all criteria.

**The ASI must be:**
- **Single**: one fix, not a list
- **Specific**: not "improve clarity" but "the second paragraph assumes the reader knows what X is — define it or move the definition earlier"
- **Concrete**: the generator should know exactly what to change
- **Actionable**: something that can be done in one editing pass

The ASI can reference any aspect of the artifact. For text artifacts it's a text instruction. If the artifact produces renderable output (SVG, code that generates images, etc.) the ASI can request the generator produce rendered output for evaluation.

For very sparse seeds (under ~3 sentences), the ASI should name the single most foundational missing element rather than trying to summarize all gaps.

## Required Output Format

```
ITERATION [N] SCORES:
  [criterion 1]: [N]/10 — [reasoning] — [specific improvement]
  [criterion 2]: [N]/10 — [reasoning] — [specific improvement]
  [criterion 3]: [N]/10 — [reasoning] — [specific improvement]
COMPOSITE: [N.N]/10

ASI (single most important fix):
[concrete, specific, actionable instruction for the generator]
```

**CRITICAL:** Use this exact format. The orchestrator and reflect subskill parse it.

## Common Mistakes

**Producing a multi-paragraph ASI**
- Problem: Generator loses focus, tries to address multiple things
- Fix: ASI is ONE sentence or short paragraph — the single highest-leverage fix

**Vague ASI**
- Problem: "Improve the tone" gives generator nothing to work with
- Fix: "The third sentence reads as condescending because of 'obviously' — reframe as an invitation"

**Anchoring to imagined intermediate scores**
- Problem: You don't have intermediate iteration scores — if you guess, you bias your judgment
- Fix: Score against the criterion descriptions and seed reference only

**Treating seed scores as a floor**
- Problem: Judge never scores below the seed, even when candidate regressed
- Fix: The seed is a calibration reference, not a minimum — score honestly

**Scoring non-integers or using half points**
- Problem: False precision, inconsistent parsing
- Fix: Integer scores only, 1-10

**Listing multiple fixes in ASI**
- Problem: Dilutes generator focus
- Fix: Pick the ONE thing that would improve the composite score the most