promptfoo/promptfoo

This page introduces an external open-source repository. It is not an HDATF product.

What it does

A CLI and library for evaluating and red-teaming LLM apps. It tests prompts, agents and RAG setups, compares models side by side, scans for vulnerabilities and runs checks in CI/CD from declarative configs.

How it helps ATF

Could be a reference for testing prompts and agent outputs in Harness and the ATF Works AI assistant in a repeatable way. Its red-team checks are worth comparing for AI features that must stay within each person's permissions.

License

MIT Permissive. Commercial use and changes are allowed if the copyright notice is kept.

More in this category

• NVIDIA/NemoClaw
  NemoClaw is an NVIDIA reference stack for running AI agents such as OpenClaw, Hermes and LangChain Deep Agents inside NVIDIA OpenShell sandboxes. Its CLI handles onboarding, managed inference, network policy, snapshots and lifecycle operations.
• vxcontrol/pentagi
  An autonomous AI agent system that carries out penetration testing tasks. It works with several LLM providers and includes agent supervision, Langfuse integration and a Graphiti knowledge graph.
• sherlock-project/sherlock
  Sherlock is a CLI tool that checks more than 400 social networks for accounts matching one or more usernames. Accounts found are saved to a text file for each username.
• QuipNetwork/hashsigs-rs
  A Rust workspace of hash-based post-quantum signature schemes, including WOTS+, SPHINCS+C and the hybrid SHRINCS signer and verifier. It also includes WebAssembly bindings and a verify-only Solana program.
• google/magika
  Magika identifies a file's content type with a small deep learning model that runs on a single CPU. It was trained on more than 200 binary and text content types and has a command-line tool.

Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.