Multi-Agent Failure Behaviors

This book started with a bug report that couldn't be filed anywhere, because nothing had technically failed. Eight subagents, each assigned a chapter of an audit plan, were asked to review their own chapter. All eight quietly answered a different question than the one asked, in exactly the same wrong way. No exception, no malformed output. Just fluent, professional, wrong. Multi-Agent Failure Behaviors is a field guide to the systematic ways LLM-based agent systems go wrong, organized into five families: Role Contamination, Underspecification, Judgment Bias, Chained Multi-Agent System Pathologies, and Goal Misalignment. Every failure behavior gets the same treatment: a plain-language definition, a narrative example that makes the mechanism intuitive, a minimal TypeScript implementation that reproduces the failure, a corrected implementation with the design principle spelled out, and a diagnostic probe you can run in CI. The thesis is simple. Most hallucinations in agentic pipelines are not invented at a single point; they are manufactured by the system. An agent fills a gap with a plausible guess, a judge waves it through, the next step strips the hedging, and three steps later it's an established fact in an executive summary. This book shows where that process can be interrupted, with defenses that don't depend on the model's cooperation: separated identities, disjoint contexts, structured data at boundaries, least privilege, mechanical enforcement. For engineers and architects building agentic systems who want a vocabulary and a checklist for failures that don't look like failures. Code samples are TypeScript against a standard LLM messages API; the mechanisms are model-agnostic. The author is a CTO, AWS AI Hero, and Cursor Ambassador who has been working on neural networks since his 2010 PhD, and on production agentic systems since they existed. Every failure behavior in this book was observed, diagnosed, and fixed in real systems before it was given a name.