Anaconda retires legacy tools in enterprise SI platform shift

Cloud package-security customers are moving to a platform with separate licensing, a platform fee and consumption charges.

Cloud package-security customers are moving to a platform with separate licensing, a platform fee and consumption charges.
Anaconda is retiring its legacy workbench and package-security products and moving cloud package-security customers to its expanded enterprise SI platform, SiliconANGLE reported on October 6. The platform has a separate license from the legacy tools, with a platform fee plus consumption charges.
The transition extends Anaconda’s business beyond distributing Python software packages. It draws on three recent acquisitions: coding company Kilo Code, security specialist Enkrypt AI and workflow provider Outerbounds. David DeSanto, Anaconda’s chief executive, said the expansion addresses business concerns about coordinating autonomous work, controlling costs and understanding agents’ access to software and data.
Anaconda already distinguishes public package access from subscription-based security services. Its support documentation says the Premium Repository requires an active subscription and an organization token, an access credential. It adds known-vulnerability scoring and package-access policy filtering that the public repository does not provide.
Kilo and Enkrypt pricing will remain unchanged until integration is complete. Migration deadlines, exact customer costs and the previous licensing terms are not yet clear.
The release includes coordination tools for “agent swarms”—groups of SI agents that carry out tasks together. Anaconda describes a lead agent assigning parts of a project to subagents that can work simultaneously, exchange information and choose different models for different jobs.
Anaconda says logs and a shared message board let customers reconstruct those interactions and inspect file access or configuration changes. That gives customers a record of what agents did, though DeSanto said the logs can run to thousands of pages. The company is working to make those records easier to digest.
Security tools acquired with Enkrypt include autonomous red-teaming agents: systems that simulate attacks to find weaknesses. Anaconda says they challenge models, agents and Model Context Protocol connections across more than 300 attack categories. Model Context Protocol, or MCP, connects agents to outside tools and information.
DeSanto said customers can use the tests to develop guardrails—rules that limit unwanted behavior—and repeat the testing after applications enter production. Anaconda also describes runtime controls that can approve, modify or block risky behavior while applications operate. The company initially gives agents no access permissions, an approach it calls least privilege.
Outerbounds technology is fully integrated into the platform. Anaconda says it supports repeatable workflows and reproducible environments, so teams can recreate the software setup used to run an application. The platform carries packages, models, and security and licensing policies from development into production rather than treating deployment as a separate setup.
Enkrypt security scores already appear on model cards, the information sheets that accompany models. Additional guardrails, combined model catalogs and enterprise controls remain on Anaconda’s roadmap. DeSanto forecast that customers would be able to buy a fully integrated package in early 2027.

Orangetown and Clarkstown have different six-month restrictions. DataBank’s expansion is exempt from the town pause but faces a disputed state decision and further environmental review.
3 min

Enterprise customers can request priority reviews of software dependencies and receive fixes adapted to the older versions they still run.

The rebuilt platform connects slide creation to company information and gives users more control over layouts and individual elements.
2 min