---
title: Trait Evaluation
description: Test who an agent is — personality consistency across interactions.
---

**Trait evaluation** tests whether an agent maintains consistent personality characteristics across different interactions and contexts.

## What it tests

- Does the agent maintain the same personality across conversations?
- Are declared traits (e.g., thoroughness, patience) reflected in behavior?
- Does personality remain consistent under different phrasings of the same question?

## Scorer: `trait_consistency`

The `trait_consistency` scorer evaluates whether the agent's responses reflect its declared personality traits:

```go
testcase.ScorerConfig{
    Name: "trait_consistency",
    Config: map[string]any{
        "trait_name":    "thoroughness",
        "dimension":     "depth",
        "expected_value": 0.85,
        "tolerance":     0.15,
    },
}
```

## Scenario: `trait_probe`

The `trait_probe` scenario generator creates multiple phrasings of the same question to test personality consistency:

```go
Case{
    ScenarioType: testcase.ScenarioTraitProbe,
    Input:        "How would you approach reviewing this code?",
    Context: map[string]any{
        "trait":     "thoroughness",
        "dimension": "depth",
        "expected":  "exhaustive analysis",
    },
}
```

## Dimension score

```go
result.DimensionScores["trait"] // 0.0 to 1.0
```

## Use cases

- Verify a thorough agent provides detailed analysis consistently
- Test that a patient agent doesn't rush through complex queries
- Ensure a cautious agent consistently identifies risks