New MCP specification kills old risks but opens fresh attack surfaces, Akamai finds
What happened
A significant update to the Model Context Protocol (MCP) due next month removes longstanding protocol-level security vulnerabilities. Akamai Technologies’ analysis of the MCP 2026-07-28 specification shows the biggest architectural change since Anthropic PBC created the original standard. While the new MCP eliminates old risks, it introduces new attack surfaces for developers and operators to address.
Why it matters
Removing old security flaws strengthens trust in the MCP and lowers the risk of exploitation through those known weaknesses. However, the fresh attack surfaces shift the defensive burden onto implementers. Builders and infrastructure operators now face new challenges protecting against threats previously unseen. This change forces a re-evaluation of security assumptions built into tooling, automation, and integration workflows using MCP.
The update reshapes where risk concentrates. It weakens legacy threat vectors but raises costs of security research and defensive efforts on new ones. For investors and businesses relying on MCP-based systems, the change pressures teams to update their threat models and invest in fresh safeguards, or potentially suffer breaches from novel exploits.
What to watch next
Focus will be on how quickly developers respond to the new attack surfaces and whether vendors release patches or tools addressing them effectively. Early adversary activity probing these fresh MCP vectors will also be a critical signal of real-world risk. Monitoring network activity and incident reports in the wake of the MCP update will show whether the specification change strengthens overall security or merely shifts the target.
Expect security teams to adjust best practices and conduct audits focused on the new vectors rather than legacy ones. The transition period will test operators’ ability to adapt and patch fast to avoid exploits. Meanwhile, regulators and compliance regimes may tighten scrutiny on MCP implementations to ensure the new risks are managed properly.
AI Quick Briefs Editorial Desk