Skip to main content

raxIT Labs · The Knowledge Base

The AI Security Knowledge Base

Real AI incidents, threats, controls and compliance clauses. The same source raxIT's agents ground their assessments in, open for anyone to browse.

Browse the Knowledge Base

4,300+ incidents2,400+ threats100+ controls300+ clauses200+ vocabulary terms

The landscape

The same knowledge base our agents use.

raxIT's specialized agents run risk assessments, threat models, control mappings, and compliance checks. Every one of them grounds its output in this knowledge base: documented incidents and controls instead of model guesswork.

We are opening it to everyone. The source our agents cite is one you can browse, query, and verify yourself.

Risk AssessmentsThreat ModelingControl MappingCompliance Checks

  1. Collect

    Automated collectors pull incident reports from 10+ public sources.

  2. Enrich

    Severity, CVSS, framework crosswalks, and a quality score per record.

  3. Publish

    Quality-gated records land on the open, CORS-readable API.

  4. Consume

    Browse here, query the API, or wire it into an agent over MCP.

Real incidents, not synthetic examples.

Every record traces back to a documented AI incident from sources like AVID, AIID, NVD, MITRE ATLAS, arXiv, and vendor disclosures.

Each one carries severity, CVSS scoring, affected systems, a quality score, and crosswalk references to OWASP LLM Top 10, MITRE ATLAS, and NIST AI RMF, so a single record answers which control gap it points to.

API and MCP access

The JSON API at kb.raxit.ai/api/v1 is open for reads: anonymous, CORS enabled, no key required.

MCP server and SDK access, for wiring the knowledge base directly into an agent or workflow, is granted per partner.