nprofile1qy2hwumn8ghj7un9d3shjtnddaehgu3wwp6kyqpqf6a33pfyp67y8llhunlhrf855xm47n3fdqymvxfj7yx78c6vqf4sy8ssyg (nprofile…ssyg) This is an interesting topic, but it feels very akin to third-party code being untrusted by default. I'm not sure how MCP is any more risky than installing a third-party package into your Django app.
I don't think it was feasible for most companies to review/audit every line of code that changes related to their stack, but I suspect we already have AI security review companies who will just throw more AI at the problem.