Cognous unveils Open Control Stack to audit AI agent decisions in minutes
Cognous has developed an Open Control Stack designed to provide rapid auditing for AI agent incidents. The system allows CISOs to review what an AI agent actually did during a specific run against defined policies. It organizes governance into four layers: declaring permitted actions, controlling runtime behavior, replaying runs, and packaging evidence. The platform integrates with LangChain to record decisions and intercept unauthorized tool calls. The resulting data can be compiled into a portable ReplayBundle for external review.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in