AI Agent Infrastructure: Compute, Protocols & Rails

Infrastructure

Test Endpoints, Unauthenticated RCE: How a LiteLLM Preview Feature Became a CVSS 10.0 Attack Vector

A command injection flaw in LiteLLM's MCP test endpoints, chained with a Starlette authentication bypass, reveals how developer convenience features are becoming the weakest link in agent infrastructure security.

◆ Blair HayesJul 12

Infrastructure

Agent Orchestration Platforms Are Federal Infrastructure Risk — Why CVSS Scores Alone Mislead Defenders

A CVSS 9.9 IDOR flaw in the agent orchestration framework reveals how attackers actually prioritize exploits – and why CVSS scores alone mislead defenders.

◆ Blair HayesJul 12

Infrastructure

Cloudflare Ships the Complete Agent Runtime Stack — and Proposes ‘Cloud 2.0’

Cloudflare's Agents Week delivers execution, memory, security, and observability in a single vendor ecosystem — the most comprehensive agent infrastructure launch from a major cloud provider to date.

◆ Blair HayesJul 12

Infrastructure

Workato Makes Agent Infrastructure Headless — Embeds Governed AI Agents Into Any Application Surface

The headless pattern that transformed CMS and commerce is reaching agent infrastructure, as Workato decouples agent logic from chat UIs and bakes governance into the transport layer.

◆ Blair HayesJul 10

Infrastructure

Google Open-Sources Agent Executor and Agent Substrate — Building the Runtime Layer Under Agent Frameworks

Google's open-source releases formalize the Execute and Constrain pillars of agent infrastructure, with Kubernetes becoming the de facto substrate for agent workloads.

◆ Blair HayesJul 10

Infrastructure

The Agent Runtime Crystallizes Into Four Pillars — Execute, Constrain, Observe, Improve

As enterprise agent adoption scales, infrastructure is formalizing into a modular four-pillar architecture that's replacing monolithic platforms with specialized, best-of-breed components.

◆ Blair HayesJul 10

Infrastructure

MCP Formalizes Extensions Framework — The Protocol Becomes a Platform

SEP-2133 introduces reverse-DNS identifiers, official and experimental extension tracks, and capability negotiation, transforming the agent protocol into a modular platform with a governed extensibility layer.

◆ Blair HayesJul 10

Infrastructure

Google ADK v2.4.0 Lets Agents Compose Entire Workflows as Tools

Version 2.4.0 introduces Workflow as Tool, ManagedAgent, mTLS, and Daytona integration — four layers that make ADK the deepest vertical integration play in the agent SDK space, and the deepest lock-in risk.

◆ Blair HayesJul 10

Infrastructure

Cisco Proposes Reusing AirPlay-Style Discovery for AI Agents — No New Protocols Required

An IETF draft from Cisco Systems defines a single '_agent._tcp' service type over mDNS/DNS-SD that lets agents find each other on local networks with zero configuration. A2A is the first protocol instantiation.

◆ Blair HayesJul 10

Infrastructure

IETF Publishes First Map of Agent Discovery Fragmentation — 15 Mechanisms, No Standard

Ericsson researchers survey 15+ competing discovery approaches across IETF, AAIF, 3GPP, and open-source. Eight gaps identified. No single mechanism spans all use cases.

◆ Blair HayesJul 10