Perplexity open-sources Numbat to monitor and block risky AI agents
The endpoint tool watches coding agents across desktop, CLI and IDE environments, but its shipped rules default to monitoring rather than enforcement.
By Ryan Merket · Published
Primary source: X
Why it matters
Coding agents can read files, run commands and touch credentials. Numbat puts detection at the endpoint where those actions occur and exposes Perplexity's internal control layer.
Perplexity, the AI search company led by co-founder Aravind Srinivas (@AravSrinivas), open-sourced Numbat on July 29th, giving security teams a common monitoring and enforcement layer for AI agents running on employee computers. Perplexity announced the release in a thread on X and published the code under an Apache 2.0 license.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.