# Agent Harness Glossary

> Definitions for agent harness, harness engineering, agent observability, agent evals, guardrails, and trajectory evaluation.

- HTML: https://openlit.io/glossary
- Markdown: https://openlit.io/glossary.md

## Terms

- [Agent harness](https://openlit.io/glossary/agent-harness.md): An agent harness is everything in an AI agent except the model: the tools, context, prompts, memory, hooks, guardrails, and feedback loops that turn a model into a working agent.
- [Harness engineering](https://openlit.io/glossary/harness-engineering.md): Harness engineering (or agent harness engineering) is the discipline of designing, measuring, and improving everything around the model in an AI agent so the agent is reliable in production.
- [Agent observability](https://openlit.io/glossary/agent-observability.md): Agent observability is the practice of tracing and monitoring the full agent harness: LLM calls, tool calls, MCP requests, retrieval, multi-step trajectories, cost, latency, and outcomes.
- [Agent evals](https://openlit.io/glossary/agent-evals.md): Agent evals are systematic checks that score agent quality, safety, and cost on real traces or offline datasets—using LLM-as-a-judge, programmatic rules, or human review.
- [Guardrails](https://openlit.io/glossary/guardrails.md): LLM guardrails are runtime controls that detect or block unsafe prompts and responses—such as prompt injection, sensitive topics, and topic restriction—before they reach users or tools.
- [Trajectory evaluation](https://openlit.io/glossary/trajectory-evaluation.md): Trajectory evaluation (or tool-call evaluation) scores the sequence of steps an agent took—tools chosen, arguments, ordering, and intermediate decisions—not only the final response.

## Related

- Pillar: https://openlit.io/agent-harness-engineering.md