# Agent Trace Evaluation

A gallery of labeling and evaluation interfaces built with Label Studio Enterprise programmable interfaces. Each entry is a starting point to adapt, not a fixed template.

**Category:** LLM & Agent Evaluation  
**URL:** https://humansignal.com/use-cases/agent-trace-evaluation

Read a multi-step LLM agent trace, rate each step correct, partial, or incorrect, score the run against a rubric, tag failure modes, and set an overall verdict, each saved as its own result.

An agent run is not one answer to grade. It is reasoning, tool calls, subagents, artifacts, and errors, and the failure is usually somewhere in the middle. This interface lays the whole trace out so a reviewer can find it.

Steps render in order with their arguments, results, tokens, and latency, and a minimap sized by latency, tokens, or cost shows where the run spent its budget. Reviewers rate each step, set a verdict, score a configurable rubric (helpfulness, faithfulness, efficiency, tool use, and instruction following by default), tag failure modes like hallucination or wrong tool, and write a critique.

Each section saves independently, so partial reviews are fine. Verdicts, step ratings, failure modes, and rubric items are all parameters.

## More LLM & Agent Evaluation interfaces
- [Multi-Turn Conversation Evaluation](https://humansignal.com/use-cases/multi-turn-conversation-evaluation)

[All use cases](https://humansignal.com/use-cases)