Three AI coding agents leaked secrets through a single prompt injection. One vendor’s system card predicted it. Prompt injection exfiltrates secrets across multiple coding agents, exposing runtime policy and isolation failures and validating the need for system cards and documented operational constraints — critical for Principle 13 (Documentation) and Principle 14 (Immune System).

CrabTrap: An LLM-as-a-judge HTTP proxy to secure agents in production. Brex demonstrates a practical runtime control that judges and blocks agent requests in real time, giving outcome engineers a deployable pattern for policy enforcement and request-level auditing (Principle 14).

Boomi builds a role for agents and guardrails in the data-connected enterprise. Boomi embeds agent-based automation with governance guardrails across data workflows, showing how orchestration and access controls scale agentic workflows safely in production (Principle 09).

What AI model should you use for revenue intelligence? Von says all the big ones, and it will automate mixing and matching for you. Von builds a context graph and dynamically routes work across Claude, ChatGPT, and Gemini, illustrating multi-model routing and context-aware model selection you can copy into outcome pipelines (Principles 06 and 11).

GoModel — open-source AI gateway in Go (44x lighter than LiteLLM). GoModel offers a lightweight, OpenAI-compatible gateway to unify multiple LLM providers, lowering integration friction for multi-provider deployments and giving engineers a simple ops layer for model switching and resilience (Principles 06 and 09).