DeepTeam
open-sourceSummary
DeepTeam is an open-source red teaming framework for LLM systems, distributed free under its published licence, and used as a library imported into your own code rather than run as a hosted service. It simulates attacks — jailbreaking, prompt injection, multi-turn exploitation — against agents, RAG pipelines, and chatbots to surface issues like bias, PII leakage, and SQL injection, and it also offers guardrails to catch these in production. It runs locally on your machine and is built on DeepEval, the project's own open-source LLM evaluation framework, which places it in the same lineage rather than alongside a named external competitor. It's aimed at teams building or securing LLM applications who need to test them the way penetration testers approach conventional software. With 32 contributors and 701 commits in the past year across four tracked releases, it is under active development.
What DeepTeam answers
Which vulnerability types does it check for beyond what a team would think to test manually?
it ships a library of attack simulations covering bias, PII leakage, and SQL injection rather than requiring each scenario to be hand-written
Does testing require calling out to a hosted service?
it runs locally on your machine as a library you import, so the attack simulations execute inside your own codebase
Can it also help outside of a one-time test run?
it provides guardrails meant to sit in production and catch the same issues after deployment, not just during a pre-release scan
What kind of systems can it actually target — just a chat endpoint?
it's built to run attacks against agents, RAG pipelines, and chatbots, not only a single prompt-response model
Do I need to build my own evaluation layer to judge whether an attack succeeded?
it's built on the project's own evaluation framework, so scoring attack outcomes reuses that existing machinery instead of a separate one
Is this a stable, finished tool or still actively changing under me?
with 32 contributors and 701 commits in the past year, the attack and vulnerability library is likely to shift release to release rather than stay fixed
Examples
Command line
No option matches that search.
| option | found in | since | description |
|---|
No option matches that search.
Values are placeholders taken from each option’s declared default. Nothing is executed here — the output shown is a recording of a run that already happened.
Release history
Nothing shipped since we started watching. Releases and docs updates land here as they happen.