AI Security
Giskard Hub
Red-teams LLM, RAG, and agent apps for prompt injection, jailbreaks, and quality failures.
Giskard Hub Overview
What it does
Giskard is an AI red teaming platform that continuously tests large language model (LLM) applications, retrieval-augmented generation (RAG) systems, and conversational agents for security and quality failures before they reach production. It uses a black-box approach that needs only API access, running a library of more than 50 adversarial probes to surface vulnerabilities a development team would otherwise miss. It is offered as an open-source library and an enterprise Hub.
How it works
The platform generates adversarial tests from security taxonomies, external threat resources, and a customer's own knowledge base, probing for prompt injection, data disclosure, hallucinations, and harmful content. Tests run from a user interface or a Python software development kit (SDK) and integrate into CI/CD pipelines, comparing model versions to catch regressions and enriching test datasets as new vulnerabilities are found. Results map to frameworks including the OWASP LLM Top 10 and the EU AI Act for audit-ready reporting.
Credentials and traction
Giskard Hub is SOC 2 Type II certified, with GDPR and HIPAA compliance for handling data in regulated environments. It is used by enterprise AI teams at Michelin, BNP Paribas, and Decathlon, and is aimed at organizations in banking, insurance, and other regulated sectors deploying generative AI applications.
Key Capabilities
mapped to solution categoriesAutonomously plans and executes multi-step adversarial campaigns against AI systems, emulating real attacker workflows across reconnaissance, exploitation, and escalation rather than running a fixed checklist of tests.
Tests LLMs and AI applications against a library of direct and indirect prompt-injection and jailbreak techniques, reporting which payloads bypass system instructions and safety controls.
Re-runs red-team campaigns continuously and at release gates in the CI/CD pipeline as models, prompts, and configurations change, catching new exploit paths before and after deployment.
Reports validated AI vulnerabilities with reproduction evidence, attacker context, and remediation guidance, mapped to the OWASP LLM Top 10, MITRE ATLAS, EU AI Act, and NIST AI RMF for auditable AI risk reporting.
Attacks deployed guardrails, system prompts, and content filters to measure how reliably they block adversarial inputs, quantifying bypass rates rather than assuming the controls work.
Routes high-value automated findings to specialist AI red teamers for manual exploitation, chaining, and depth beyond automated coverage, blending platform testing with human expertise.
Compliance
certificationsImplementation & support
Info last updated on July 11, 2026
Vendors
Is this your product?
Claim your profile to connect with the teams looking for your solutions.