Anaconda has unveiled new platform capabilities that combine agent swarms with autonomous security testing, aiming to help enterprises secure AI systems as they build. The company says the expansion pairs agentic development with red-team agents, trusted packages, and governance controls.
The move comes as adoption of multi-agent systems grows. Anaconda's survey of AI-native builders found 63% are moving toward agent swarms. Separately, Enkrypt AI research cited by the company scanned more than 268,000 agent tools across over 25,000 MCP servers and found vulnerabilities in 73% of them. The company also cited Omdia data that 72% of organizations rank managing agent autonomy as critical.
New capabilities include AI Workspaces, VS Code integration for agent swarms via Kilo, and Kilo Desktop for local development with access to hundreds of models and thousands of vetted packages. On the security side, Anaconda is introducing red-teaming across more than 300 attack categories, runtime guardrails, and an Agent Incident Registry that provides a source-backed record of publicly reported agent incidents.
The announcement is based on a single source, a press release from Anaconda. It includes quotes from Anaconda CEO David DeSanto and Omdia analyst Mark Beccue, as well as a customer testimonial from IntraPhone's Robert Björne. No independent verification of the cited statistics was provided.