Jeffrey Aven
Maintainer
StackQL Studios
Query before mutation: guardrail patterns for agents that touch production infrastructure
Tuesday, 8 December · 12:40 PM · AI Engineering track
Query before mutation: guardrail patterns for agents that touch production infrastructure
Giving an AI agent permission to change production infrastructure removes the human pause that made plan-and-apply safe. In this talk, Jeffrey Aven shows how to put that judgement back into the system itself.
He introduces query before mutation: requiring an agent to establish the live state of an environment before it can write. From there, he develops a practical set of guardrails: idempotent assertions instead of imperative commands, machine-checkable policy gates, tightly scoped credentials and mutation surfaces, and recovery paths for blocked or failed changes.
Drawing on real examples of agents operating against cloud environments, Jeffrey will show where each pattern helps, where it does not, and demonstrate a mutation being blocked and safely recovered.
You’ll leave with an ordered checklist of controls that deliver the most protection for the least implementation effort.
Jeffrey Aven
Jeffrey Aven is a software and systems architect, engineer, advisor, educator, author and founder with 35 years of experience in the technology industry. He has written several books on data and analytics and is an authorised trainer for technology vendors including Google, Confluent and Databricks.
Jeffrey is a member of the Google Developer Expert (GDE) program, Google’s invitation-only developer advocate community, and has extensive experience and accreditation across Anthropic and Claude, AWS, Snowflake, Databricks and other cloud, data and SaaS platforms.