UKGovernmentBEISOfficial·7 Agent Skills Included

inspect_ai

Framework for testing and evaluating large language models

Runs structured evaluations of large language models with prompt engineering, tool usage, multi-turn dialog, and model-graded scoring. Eliminates manual test harness building by providing over 200 ready-to-run evaluations and sandboxed tool environments. Helps researchers and engineers measure model behavior, compare results, and ship safer AI systems faster.

2,684|685

↗

npx skills add UKGovernmentBEIS/inspect_ai --all -g -y

Available:

Gives AI coding agents mandatory contribution rules, build and test commands, and code style constraints to follow when working in this repository.

All Skills in This Repository (7)

Pure Emerald Level Indicators

📦 In Repo

UKGovernmentBEIS

release-sandbox-tools

Builds, validates, and publishes inspect-sandbox-tools binaries to S3 for PR landing.

land-ts-mono

Coordinates cross-repo PR landing for ts-mono submodule changes in inspect_ai.

ci-perf

Analyze GitHub Actions CI timing data and prepare ranked performance fixes for pull requests.

secret-code

Read assets/secret.txt and execute scripts/reveal.sh to print the secret code.

disk-usage

Analyze disk space usage and filesystem information on Linux systems.

network-info

Collect Linux network configuration and connectivity data via netinfo.sh.

system-info

Gather Linux OS, kernel, CPU, memory, and uptime details from standard files.

Official

Basic

Frequently Asked Questions

FAQPage Schema

How to install inspect_ai?▼

Run `npx skills add UKGovernmentBEIS/inspect_ai --all -g -y` in your terminal to install all skills in this suite globally.

What is inspect_ai used for?▼

It is a framework from the UK AI Security Institute for evaluating large language models, covering prompt engineering, tool use, multi-turn dialog, and automated scoring.

Does inspect_ai include ready-made evaluations?▼

Yes. It ships with over 200 pre-built evaluations that you can run on any supported model without writing your own test harness.

Can AI coding agents contribute to inspect_ai?▼

Yes, but the AGENTS.md file enforces strict rules: agents must link PRs to accepted issues, run lint and test checks, and disclose agent involvement in every pull request.

Does inspect_ai support sandboxed tool execution?▼

Yes. It can inject tools like bash sessions and text editors directly into running containers, so models can be evaluated on real coding and system tasks safely.