Anaconda Buys Enkrypt AI, Citing 143,000 MCP Flaws

Anaconda acquired Enkrypt AI on August 4, per the company's announcement. The deal folds AI red-teaming and compliance tools into a company best known for distributing Python packages.

Enkrypt's scans found 143,000 vulnerabilities across 73% of the MCP servers it checked. That's the number worth sitting with if your team has wired an LLM into internal tools over MCP and never audited the connection.

The acquired technology covers three things: pre-deployment red-teaming across more than 300 attack categories. It also adds runtime guardrails against jailbreaks and data leaks, plus compliance mapping to the NIST AI RMF and the EU AI Act. The EU AI Act's obligations took effect August 2, two days before the deal closed.

This is Anaconda's second acquisition in two months, following Kilo Code. Two deals that close together is a pivot, not a side bet. A company built on packaging open-source Python libraries is now selling AI governance tooling on top of that distribution business.

The take

That bet holds only if the 73% failure rate survives scrutiny from teams auditing their own servers. Enkrypt's numbers now come from a company with a financial reason to keep finding vulnerabilities. I'd watch whether Anaconda folds the scanning into its core distribution product or sells it as a separate layer.