Search across all documentation pages
11 pages in this section.
Learn how prompt injection, tool access, and retrieval combine into real security risks in Claude applications. Understand key concepts and mitigation strategies.
Learn to build secure RAG pipelines with 8 examples. Discover how to keep secrets out of prompts, validate inputs, and prevent data exfiltration.
Learn to defend against indirect prompt injection in RAG pipelines. Sanitize and isolate untrusted retrieved content to prevent hijacking Claude's instructions or tools.
Restrict tool permissions for Claude agents before deployment. Learn to scope tools to minimum necessary capabilities, identify data access, and prevent prompt injection
Understand how to redact, retain, and audit PII in Claude API prompts and logs to maintain SOC2 and GDPR compliance.
Reduce RAG hallucinations with this checklist of citation and grounding techniques. Learn to keep Claude's answers tied to retrieved source documents.
Learn best practices for securing Claude applications using tools and sensitive data. Understand threat modeling, injection defense, and retrieval quality.
A single-page roundup of every highlight bullet from the 10 pages in the Security & RAG section, grouped by source page so you can scan all 60 takeaways without opening each article individually.