UKGovernmentBEIS/inspect_ai ? reverse-engineered prompt

Build me a Python framework for evaluating large language models.

I want a tool that lets me define evals, run prompts through models, support tool use and multi turn conversations, and score the results automatically, including model graded judging. It should come with a bunch of ready to run example evaluations so I can test different models right away, and it should be easy to add my own custom evals, scoring rules, and extensions later.

Please make it feel like a real developer friendly library, with clear docs, a simple command line workflow, and a small web UI if that fits naturally. I’d also like the project set up so it’s easy to install in editable mode for local development, with tests, linting, formatting, and a clean path for contributors to work on it. If you need to check current docs or best practices while building, look them up online.