garak
Named by 12 kept items in this workspace.
11 items name it under tools and 1 under classification.technologies.
The items that name it 12
Referenced as a real tool, an LLM vulnerability scanner used in the labs.
Removed from list on 2026-09-07 (pruned)
a LLM vulnerability scanner
Generative AI red-teaming & assessment kit / LLM vulnerability scanner that probes LLMs for hallucination, data leakage, prompt injection, misinformation, toxicity, jailbreaks and other weaknesses; analogous to nmap/Metasploit but for LLMs
Automated LLM vulnerability scanner; now maintained by NVIDIA; v0.14.x in development June 2026 adding probes for agentic AI systems; 50+ specialized probes
Open-source LLM vulnerability scanner; probes live inference endpoints for prompt injection (AML.T0051), jailbreak susceptibility (AML.T0054), data leakage, hallucination under adversarial inputs; does not evaluate cloud infrastructure or IAM permissions
LLM vulnerability scanner (--model_type usage referenced); bundled probe payload datasets
Generative AI Red-teaming and Assessment Kit — LLM vulnerability scanner with 100+ attack modules, described as 'functions like nmap but for LLMs'; supports hallucination detection, data leakage assessment, prompt injection testing, toxicity/bias evaluation, jailbreak analysis
Integrated as Web Injection and Doctor scenarios for LLM vulnerability scanning
LLM vulnerability scanner; this is the v0.17.0 release covering new features, plugin fixes, and tuning
LLM vulnerability scanner / red-teaming framework by NVIDIA, releases tracked v0.13.0 through v0.17.0 with probes, detectors, and generators for testing target LLM systems
In every workspace you can read 13
Security
All technologies in this workspace.