Agent Infra: Standards, Search, Serverless, Channels, Evals

OpenAI introduces Agent Plugins open standard for bundling skills and MCP servers; steering committee includes Amazon, Microsoft, Vercel — OpenAI publishes an open Agent Plugins standard and steering committee to bundle skills and MCP servers into interoperable packages. Outcome engineers get a de facto packaging and discovery format to standardize skill deployment and orchestration across providers, reducing integration friction and shaping governance boundaries (Principle 09).

Cloudflare AI Search: give your agents a search engine for your data — Cloudflare launches AI Search that provides agents with built-in semantic search across sites and docs plus easy ingestion and public endpoints. This lets you replace brittle scraping with reliable retrieval and predictable pricing, tightening context engineering and making grounding scalable (Principle 06, Principle 11).

The next generation of MCP — Cloudflare updates MCP to a stateless design, enabling serverless Workers hosting and removing sticky-session complexity for agent integrations. Stateless MCP lowers operational cost and simplifies edge-hosted agents, so you can run distributed, low-latency toolchains without session plumbing (Principle 06, Principle 11).

Channels SDK — Bring Any Agent to Any Channel (Slack, Microsoft Teams) — CopilotKit ships a Channels SDK that embeds AG-UI agents into Slack and Teams with native UI, tool access, and human approval gates. Use this to make agents first-class in collaboration workflows, enforce human approval and access controls, and accelerate production rollout in communications platforms (Principle 09, Principle 15).

Do evals the Airbnb way — Pydantic documents an eval-driven development stack that combines programmatic checks, LLM judges, and human calibration to reproduce Airbnb’s eval workflow. This gives a practical blueprint for continuous evals and human calibration so you can measure outcome quality, catch silent failures, and keep automated agents honest (Principle 16, Principle 14).