+16
日增长
10,037
stars
1,481
forks
66
watchers
1,355
prs
1,017
issues
34
releases
README
garak, LLM vulnerability scanner Generative AI Red-teaming & Assessment Kit garak checks if an LLM-backed application can be made to fail in a way we don't want. garak probes for hallucination, data leakage, prompt injection, misinformation, toxicity generation, jailbreaks, and many other weaknesses. If you know nmap or msf / Metasploit Framework, garak does somewhat similar things to them, but for LLMs. garak focuses on ways of making an LLM or dialog system fail. It combines static, dynamic, and adaptive probes to explore this. garak's a free tool. We love developing it and are always interested in adding functionality to support applications. Get started See our user guide! docs.garak.ai Join our Discord! Project links & home: garak.ai Twitter: @garak_llm DEF CON slides! LLM support currently supports: hugging face hub generative models replicate text models openai api chat & continuation models aws bedrock foundation models litellm pretty much anything accessible via REST gguf models like llama.cpp version = 1046 .. and many more LLMs! Install: garak is a command-line tool. It's developed in Linux and OSX. Standard install with pip Just grab it from PyPI and you should be good to go: python -m pip install -U garak Install development version with pip The standard pip version of garak is updated periodically. To get a fresher version from GitHub, try: python -m pip install -U git+https://github.com/NVIDIA/garak.git@main Clone from source garak has its own dependencies. You can to install garak in its own Conda environment: conda create --name garak "python=3.11, garak needs to know what model to scan, and by default, it'll try all the probes it knows on that model, using the vulnerability detectors recommended by each probe. You can see a list of probes using: garak --list_probes To specify a generator, use the --target_type and, optionally, the --target_name options. Model type specifies a model family/interface; model name specifies the exact model to be used. The "Intro to generators" section below describes some of the generators supported. A straightforward generator family is Hugging Face models; to load one of these, set --target_type to huggingface and --target_name to the model's name on Hub (e.g. "RWKV/rwkv-4-169m-pile"). Some generators might need an API key to be set as an environment variable, and they'll let you know if they need that. garak runs all the probes by default, but you can be specific about that too. --spec probes.promptinject will use only the PromptInject framework's methods, for example. You can also specify one specific plugin instead of a plugin family by adding the plugin name after a .; for example, --spec probes.lmrc.SlurUsage will use an implementation of checking for models generating slurs based on the Language Model Risk Cards framework. For help and inspiration, find us on Twitter or discord! Examples Probe a commercial model for enco
Top issues
#1666Proposal: Prompt defense posture detectors — assess system prompt hardening18 comments · 2026-04-03#2136judge detectors: unparseable judge verdict is scored as a successful attack10 comments · 2026-08-31#579probe: separation10 comments · 2024-04-02#2445probe: retrieval access-control leaks in RAG assistants, using planted canary documents8 comments · 2026-11-08#2307promptinject: rogue string is translated in the prompt but not in `triggers`, so non-English runs under-report hits8 comments · 2026-10-13#844migrate away from conda8 comments · 2024-08-21#515probe: aart8 comments · 2024-02-27#2459test_tox_unsafe and test_tox_safe assert nothing: the prompt has no lang, so the detectors return no scores7 comments · 2026-11-12
Discussions
#2319v0.17.10 comments · 2026-10-16#2308Runs with zero prompts report SKIP 0/0 without showing the provider error."3 comments · 2026-10-13#2237Probe idea: output-contract adherence as a measurable property, separate from refusal0 comments · 2026-09-26#2215Free-tier Groq rate limits cause multi-hour runtime with no visible warning (parallel_attempts + low TPM)5 comments · 2026-09-18#2203New contributor interested in LLM security probes and testing1 comments · 2026-09-17#2171v0.17.00 comments · 2026-09-09#2119Benchmarked garak against two third-party targets, corrections welcome0 comments · 2026-08-26#2116finding-bridge: a hitlog consumer that seals hits for triage0 comments · 2026-08-25
Latest release
v0.17.1 v0.17.1 2026-10-16What's Changed New features feat(evaluators): add hit@k attack success rate metric by @stefanoamorelli in https://github.com/NVIDIA/garak/pull/2090 This new metric is currently exposed as terminal output and available per probe/detector pair in the final digest entry of the report.jsonl Add Python 3.14 support by @NK-Works in https://github.com/NVIDIA/garak/pull/2297 Improved plugins fix(detectors): match mitigation.Prefixes at the start of the response by @chiruu12 in https://github.com/NVIDIA/garak/pull/2067 fix(detectors): stop FileDetector scoring an unexamined file as a