# What is Harness?

The product code that runs a model on a task: the loop, tools, context, permissions and memory that turn single replies into finished work. Agentic loops work when a capable model meets a good harness; neither is enough alone.

## In real work

The new model helped, but the bigger jump came from the harness: the coding agent now runs the tests, reads the failures and tries again before replying.

## Related terms

- [Capability](https://prompthausapp.com/learn/capability.md)
- [Tool Use](https://prompthausapp.com/learn/tool-use.md)
- [Computer Use](https://prompthausapp.com/learn/computer-use.md)
- [Model Generation](https://prompthausapp.com/learn/model-generation.md)
- [Golden Set](https://prompthausapp.com/learn/golden-set.md)

## How is this different from Googling it?

Google gives a definition. Prompthaus pairs each term with how it shows up in standups, PRs, prompts, and design reviews. Then it quizzes you until the word sticks. One term a day on iPhone.

- Category: AI Product
- HTML: https://prompthausapp.com/learn/harness
- App Store: https://apps.apple.com/us/app/prompthaus-ai-vocabulary/id6776314407
