Anaconda Puts Guardrails on Agent Swarms
Techstrong TV Interviews
•
23m
AI agents now work side by side in swarms, and that raises new security questions. Alan Shimel talks with David DeSanto, CEO of Anaconda, about how the company brings guardrails, least privilege access and visibility to agent swarms.
From Python packages to an AI platform
DeSanto joined Anaconda as CEO a year ago after six years as chief product officer at GitLab. Anaconda started with Python packages and Conda, and it now serves more than 55 million users. Over the past year, it acquired Outerbounds for production AI, Kilo for agentic engineering and Encrypt AI for red teaming and guardrails. As a result, the platform now runs end to end, from building with agents to production.
Securing agent swarms
Anaconda's general availability release added agent swarms that run in parallel inside notebooks and IDEs. However, swarms raise real risks. So every agent runs with least privilege access, inside guardrails, and talks to other agents through a secure message board. In addition, an agent must ask for access to a specific file for a specific purpose, and a person can approve it once.
Agent swarms and the AI dev factory
DeSanto says traditional software development is giving way to the AI dev factory. Agents now write the code, so teams must test and secure it before it ships. For example, Encrypt AI scanned 25,000 MCP servers and found 143,000 vulnerabilities across 73% of them. That is why autonomous red teaming and guardrails matter as agent swarms spread.
A hopeful view of AI
Despite a rough month for security, both Shimel and DeSanto remain optimistic. DeSanto points to McGill University, which used Anaconda to find a $10 alternative to a $10,000-a-month cancer drug. Still, he urges teams to use secure open source, consider open-weights models and build security into every step.
Agent swarms can speed up development when they run safely. Explore more DevOps and AI videos and the latest Techstrong TV interviews.
For more information please visit anaconda.com
Up Next in Techstrong TV Interviews
-
Aviatrix Makes Breach Containment Fre...
Security teams cannot patch fast enough to keep up with this month's flood of vulnerabilities. Alan Shimel talks with Doug Merritt, CEO of Aviatrix, about Clearance and why breach containment is now the most reliable way to limit damage. Aviatrix is also making Clearance free for any enterprise t...
-
AI Agents Turn IT From Blocker to Ena...
Outcomes Over Pieces
AI agents are changing what IT teams spend their time on. For decades, vendors sold pieces, and customers had to stitch them together. People did much of that work by hand.Helix VP of SaaS Operations Tom Davies examines that shift with Mike Vizard. Helix recently became an ...
-
AI Coding Hits the Context Window Wall
More Code, Less Understanding
The context window has become a hidden limit on AI coding. Teams are adopting coding tools quickly. Yet many skip the step of helping AI understand the system it is changing.Blitzy Field CTO Neeraj Deshmukh examines that gap with Mike Vizard. He says expectations n...