garak
A command-line LLM vulnerability scanner that probes models and chatbots for prompt injection, data leakage, jailbreaks and other unwanted behavior.
About garak
Language model integrations need systematic tests for failures that ordinary happy-path checks do not reveal.
Who it’s for
- Security engineers assessing LLM applications
- Developers testing model behavior before deployment
When to consider it
Consider garak when you want configurable vulnerability probes across hosted APIs and local language models.
Tradeoffs & limitations
- The scanner is free and Apache-2.0 licensed, but requires Python 3.11 or newer; the latest development code is installed separately from the periodic PyPI release.
SOTA overview · Documentation-based assessment · Sources & review method
Updates
No updates shared yet.
Discussion
Newest firstAsk a question or share how you use garak.
Keep it helpful. Community rules
Loading discussion…
Sources & review method
Documentation-based assessment · Oct 14, 2026 · Prepared with AI assistance; not a hands-on benchmark.
- garak official website
- garak repository and installation
- garak package requirements
- garak user documentation
- Recent official maintenance commit (2026-10-10, UTC)
Checked by SOTA · AI-assisted documentation review. Selection advice is our assessment; verify current requirements for your deployment.
Import history & original evidence
garak official website
A command-line LLM vulnerability scanner that probes models and chatbots for prompt injection, data leakage, jailbreaks and other unwanted behavior.
Based on official pages and announcements checked on 2026-10-14. No hands-on test, performance benchmark or popularity ranking is claimed.