Overview
Enterprises need inventory, attribution, and policy enforcement as AI agents and MCP tools proliferate inside approved SaaS and developer workflows.
Timeline
- 2026-07-20: neo-security $100M stealth exit (neo-security-100m-ai-agent-control)
- Incident backdrop: openai-math-ai-sandbox-bypass, Hugging Face agent security incident
- Tooling: github-code-quality as adjacent quality/governance surface for AI-generated code
Key Players
Analysis
Category formation phase: capital clusters around control layers while labs publish containment failures. Buyers map inventory→enforcement stacks.