Back to all resources
AI Red Teaming Field Guide
A provider-neutral field reference for authorized testing of LLM applications, agents, RAG, MCP, models, and the infrastructure around them.
Publication status
Editorial and technical review in progress
- English Markdown source received
- Spanish edition follows the reviewed English source
- Branded PDF editions are not published yet
Publication map
The draft source is organized by attack surface. Every chapter is readable online with its review state kept visible.
Draft in review
OV-03
Orient the engagement
- OVPhiloCyber AI Red Teaming Field GuideSource: README.md4 min readRead draft chapter
- 00PrefaceSource: 00-preface.md5 min readRead draft chapter
- 01FoundationsSource: 01-foundations.md10 min readRead draft chapter
- 02MethodologySource: 02-methodology.md11 min readRead draft chapter
- 03ReconnaissanceSource: 03-reconnaissance.md13 min readRead draft chapter
04-13
Test the attack surfaces
- 04Attacking LLMsSource: 04-attacking-llms.md25 min readRead draft chapter
- 05Attacking AgentsSource: 05-attacking-agents.md25 min readRead draft chapter
- 06Attacking RAG SystemsSource: 06-attacking-rag.md16 min readRead draft chapter
- 07Attacking Model Context Protocol (MCP)Source: 07-attacking-mcp.md18 min readRead draft chapter
- 08Attacking EmbeddingsSource: 08-attacking-embeddings.md13 min readRead draft chapter
- 09Data AttacksSource: 09-data-attacks.md15 min readRead draft chapter
- 10Adversarial EvasionSource: 10-evasion.md14 min readRead draft chapter
- 11Supply Chain AttacksSource: 11-supply-chain.md22 min readRead draft chapter
- 12Infrastructure AttacksSource: 12-infrastructure.md14 min readRead draft chapter
- 13Privacy AttacksSource: 13-privacy-attacks.md17 min readRead draft chapter
14-15
Defend and integrate
R1-R2