AI agents act on business data and enterprise systems through the tools they call, and those tools can carry security risk. In two months, Enkrypt AI by Anaconda scanned 268,000+ tools across 25,000 Model Context Protocol (MCP) servers, uncovering 143,000+ vulnerabilities affecting 73% of those servers. Enterprise leaders are asking their teams to do more with agents, so the practical question is how to scale the value while keeping control.

Today, Anaconda joined the Open Secure AI Alliance to help address that question. For more than a decade, Anaconda has secured and governed the open-source software supply chain for enterprise data science, machine learning, and AI, and we bring that experience, along with our work in agentic development and AI security, to a shared industry effort.

An open approach to securing AI

Founded by NVIDIA and led by the Linux Foundation, the Open Secure AI Alliance brings together technology companies, researchers, and government partners to develop and share open tools, standards, and techniques for securing AI. Members include Anaconda partners NVIDIA, Microsoft, and AWS, alongside Cloudflare, Red Hat, and others across enterprise technology and cybersecurity.

The alliance’s approach centers on giving organizations defensive capabilities they can inspect, adapt, and run on infrastructure they control. Its scope spans the full agent stack, including models, the software that enables agents to take action, and the policies and controls that govern them.

AI spans many providers. Enterprise teams combine open and proprietary models, development tools, data platforms, and infrastructure. Shared security approaches can help organizations evaluate and protect those systems while preserving their ability to choose the right technology for the business.

Security has to cover the system around the model

A model assessment evaluates the model. But enterprise teams also need to understand the system around it: the packages it depends on, the tools it can reach, the data it can access, and the actions it can take.

Consider an agent that retrieves a document, changes its plan based on what it reads, and then updates a business record. Each step may look reasonable on its own. The full sequence can still take the agent outside the scope the business approved.

That is why security needs to span development and production, with evidence that controls work as agents, models, and connections change.

Anaconda plans to collaborate with fellow alliance members on three priorities:

  • Test agents across their full chain of decisions. Improve evaluation methods and red-teaming techniques to uncover failures that unfold across multiple steps.
  • Secure the tools and MCP servers agents connect to. Advance common approaches to scanning, vetting, and controlling access across the AI software supply chain.
  • Turn AI policy into evidence. Improve how organizations record agent decisions and report incidents so security and compliance teams can verify which controls ran and what they allowed or stopped.

A unified platform for building, securing, and operating AI

Teams building AI agents and applications need speed and control at the same time. Covering development, security, and operations with separate tools leaves policies inconsistent as AI moves between teams and lifecycle stages. Anaconda Platform gives teams four capabilities for that work.

Build with AI workspaces. Builders use agentic tools in Kilo by Anaconda, with model choice, trusted packages and models, and visibility into usage and cost. Kilo Enterprise adds model and provider controls, single sign-on (SSO), and audit logs to support adoption across teams.

Secure with AI security & guardrails. With Enkrypt AI by Anaconda, teams can red-team agents before release, enforce guardrails on agent tool calls in production, and keep audit-ready evidence for continuous compliance.

Operate with AI orchestration. Teams run reproducible, traceable workflows across environments as AI work moves into production.

Trust with AI artifacts. Builders start from vetted Python packages, curated models, and trusted MCP servers, with provenance and vulnerability tracking.

These capabilities work together to secure the enterprise. Vetted artifacts give builders a trusted starting point, security testing results feed into runtime protections, and reproducible workflows let teams trace what changed from experimentation through production. Teams can start with the capability that addresses their biggest source of friction and add the others as needs change.

Raising the bar together

Security for AI agents improves when the people who build, secure, and run them share what they learn. Anaconda has worked in the open-source community for more than a decade, and we are joining the alliance in that spirit. We support open standards for testing agents, vetting MCP servers, and recording agent decisions, and plan to share what we learn so the whole community can raise the bar on security and trust.

Our goal is for organizations to scale AI on their own terms, with builder freedom and enterprise control, using open tools and standards the community can inspect and build on.

Request a demo or start a conversation about AI security with our team.