#MCP and tools

Connector authorization was the biggest operational friction in rolling out MCP at companies, and EMA pushes it into standard SSO workflows that IT teams already run. If your identity provider does not yet support EMA, the trusted-issuer path covers 14 popular connectors today, so most Team and Enterprise tenants can turn this on immediately. This also sets a template other AI vendors may adopt, since EMA is an open MCP extension rather than a Claude-only mechanism.

宝玉@doteySource

Running observability inside the customer's own cloud removes a common blocker for enterprises that cannot send agent traces to third-party infrastructure. Teams evaluating agent monitoring stacks can now compare LangSmith BYOC against self-hosted alternatives without taking on their own maintenance. Availability spans the three major clouds, so the choice mostly comes down to where existing workloads already sit.

Official accountLangChain@LangChainAISource

Trusted-issuer login removes the repeated OAuth prompts that typically interrupt MCP plugin adoption, lowering the barrier for users to keep multiple tool servers connected. Developers shipping MCP integrations should implement the flow now so their plugins connect in one step. It aligns Claude's plugin experience with what users expect from mainstream app single sign-on.

claudedevs@claudedevsSource

It treats correctness, not speed, as the bottleneck for agent-driven analytics, requiring answers to carry verifiable business meaning and lineage. Data and platform teams can layer it on existing semantic models instead of rebuilding them, and expose certified metrics to any MCP-capable agent. Deployment is self-hosted via NGC Helm charts or Docker Compose, with self-service registration disabled and a bootstrap admin required for governance.

shao__meng@shao__mengSource

Version-matching Dynamo with the surrounding infrastructure is the fiddly part of running inference on Kubernetes, so a tested, validated bundle removes a real source of deployment failure. Platform and MLOps engineers provisioning GPU clusters can follow the official walkthrough to configure and validate their setup. The KubeCon North America session is the place to see it demonstrated live.

Official accountNVIDIA AI@NVIDIAAISource

The non-obvious point is that accessibility-tree automation trades some visual accuracy for speed and lower token cost, which could make long-running desktop agents far cheaper to operate. Builders experimenting with computer-use agents should test it against vision-based alternatives like Codex's computer use to see where the accuracy-versus-latency tradeoff lands for their workflows.

Viking@vikingmuteSource

The shift from listing dozens of MCP tools to writing code against them changes how agents scale with large toolsets, cutting token overhead and improving reliability. Developers building agent workflows can try the Gemini 4 Argon example or the pi framework directly. Schmid's bet that MCP remains the underlying protocol, just invoked through code, is a useful planning signal for anyone wiring up tool infrastructure.

Philipp Schmid@_philschmidSource

The shift changes how agent harnesses should be designed: instead of engineering tool schemas and prompts, expose typed functions and let the model orchestrate in code, which also cuts token costs from intermediate outputs. Builders of agent frameworks and MCP servers should consider adding exec-style sandboxes and outputSchema support. Typed JS/Python sandboxes beat bash for chaining MCP tools since models skip shell-escaping and CLI flag discovery.

shao__meng@shao__mengSource

Agent workflows that loop heavily on Search now carry a separate tool bill on top of model token costs — 10,000 successful searches a day adds roughly US$7.50 before inference fees. Developers building on Claude Code or Codex against DeepSeek endpoints should audit their call volumes and consider caching or batching to control spend. The per-account QPS caps and 402-on-low-balance behaviour also matter for anyone running production agents at scale.

shao__meng@shao__mengSource