AI & LLM
LLMIntegrationSecurityChecklist
Ship customer-facing LLM features without leaking secrets, data, or trust.
PDF28 pages20 min
The 60-item security checklist our team uses when shipping LLM-powered features into regulated environments. Covers prompt injection, data exfiltration, output handling, tenant isolation, and audit trails — mapped to OWASP LLM Top 10 and NIST AI RMF.
What's inside
- 60 controls mapped to OWASP LLM Top 10 and NIST AI RMF
- Prompt-injection test cases you can paste into your eval suite
- Tenant-isolation patterns for multi-tenant RAG systems
- Redaction, logging, and audit-trail requirements for regulated data
- Incident response runbook template for model-driven breaches
