# confident-ai/deepteam

DeepTeam is a framework to red team LLMs and AI agents.

Repository: https://github.com/confident-ai/deepteam
Canonical: https://ross.abutalabs.com/products/deepteam
Homepage: https://trydeepteam.com
Language: Python
License: Apache-2.0
License Family: permissive
Topics: llm-guardrails, llm-red-teaming, llm-safety, python, llm-seecurity
Last push: 2026-08-21T03:45:15+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 98, release rhythm 44, longevity 39
- inputs: {"age_days": 546, "days_push": 12, "days_rel": 294, "gap_med": 86.0, "n_releases_24m": 3}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2623, forks 426 (observed 2026-08-28T04:07:05.136846+00:00)

## What it is
DeepTeam is an open-source Python framework for red teaming LLM systems, AI agents, RAG pipelines, and chatbots. It simulates adversarial attacks such as jailbreaking and prompt injection to surface vulnerabilities, and maps risk assessments to frameworks like OWASP Top 10 for LLMs, NIST AI RMF, MITRE ATLAS, and the EU AI Act.

## Use cases
- red team my llm application for jailbreaks and prompt injection
- test my ai agent for security vulnerabilities before production
- run an OWASP Top 10 for LLMs risk assessment from python
- detect PII leakage and bias in my chatbot
- simulate multi-turn adversarial attacks against my RAG pipeline
- check my llm app against NIST AI RMF or EU AI Act requirements
- add guardrails to block unsafe prompts in production

## When to choose
- you need adversarial security testing for LLM apps, agents, or RAG systems
- you want framework-aligned assessments (OWASP, NIST, MITRE ATLAS, EU AI Act) from a Python API or YAML CLI
- you want a local, open-source alternative to commercial LLM pentesting tools
- you already use DeepEval and want dedicated red teaming alongside evaluation

## When to avoid
- you need general LLM quality evaluation like correctness or faithfulness - use DeepEval instead
- you need traditional network or web application penetration testing
- you need a hosted, no-code-only security platform without local execution

## Facets
- artifact type: framework
- maturity: active
- function: security, penetration-testing, testing, llm-inference, agent-framework, rag, chatbot, cli
- domain: security, large-language-models, machine-learning, developer-tools, penetration-testing
- platform: python, cli, cross-platform
- tags: llm-red-teaming, llm-safety, llm-guardrails, jailbreaking, prompt-injection, owasp-top-10, nist-ai-rmf, mitre-atlas, eu-ai-act, vulnerability-scanning, ai-security, ai-agents

## Member repositories
- confident-ai/deepteam (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:05.136846+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:19:57.858682+00:00, confidence not recorded.
  - readme: https://github.com/confident-ai/deepteam (fetched 2026-08-28T04:07:05.136846+00:00, sha 722539443331)
  - homepage: https://trydeepteam.com (fetched 2026-08-29T10:03:04.226730+00:00, sha b9c9fda56ff2)
  - site_page: https://www.trydeepteam.com/docs/getting-started (fetched 2026-08-29T10:03:04.236288+00:00, sha 793d484ceea1)
  - site_page: https://www.trydeepteam.com/docs/frameworks-owasp-top-10-for-llms (fetched 2026-08-29T10:03:04.239608+00:00, sha 41ae3c510267)
  - site_page: https://www.trydeepteam.com/docs/frameworks-mitre-atlas (fetched 2026-08-29T10:03:04.242239+00:00, sha db89ff74764e)
  - site_page: https://www.trydeepteam.com/docs/frameworks-nist-ai-rmf (fetched 2026-08-29T10:03:04.245083+00:00, sha bd367c7cd0e9)
  - site_page: https://www.trydeepteam.com/docs/frameworks-eu-ai-act (fetched 2026-08-29T10:03:04.248029+00:00, sha 727e54277c28)
  - site_page: https://www.trydeepteam.com/docs/frameworks-beavertails (fetched 2026-08-29T10:03:04.250769+00:00, sha 2763714ac836)
  - site_page: https://www.trydeepteam.com/docs/frameworks-aegis (fetched 2026-08-29T10:03:04.252620+00:00, sha e1ac783e88b5)
  - site_page: https://www.trydeepteam.com/docs/red-teaming-yaml-cli (fetched 2026-08-29T10:03:04.254775+00:00, sha c56e19014e46)
- Data as of 2026-08-30T08:39:29.467469+00:00.