MCP App StoreThe latest from the AI and MCP ecosystem, curated daily.
This week marks a significant shift from general AI capabilities toward a more structured, enterprise-ready agentic ecosystem. The dominant theme is the transition from "experimental" to "operational," characterized by new controls for agent steering, centralized authorization, and the expansion of agent logic into physical hardware.
The Model Context Protocol (MCP) took a major leap toward organizational adoption with the introduction of Enterprise-Managed Authorization. By enabling zero-touch OAuth via identity providers like Okta, MCP is removing the authentication friction that previously hindered large-scale deployments. This is paired with Anthropic's move toward Workload Identity Federation, replacing static API keys with short-lived credentials to harden security for enterprise workloads.
We are seeing a move away from simple prompt-engineering toward formal steering frameworks. Anthropic's comprehensive guide on Steering Claude Code—utilizing CLAUDE.md, skills, and subagents—provides a blueprint for making agentic coding predictable. This control is further enhanced by the integration of Artifacts in Claude Code, bringing visual feedback directly into the CLI loop.
Beyond the screen, Hugging Face's work with Strands Agents and LeRobot is bridging the gap between digital model hubs and physical robot hardware. Simultaneously, the launch of Agentic Resource Discovery allows agents to autonomously find the tools and data they need, while the release of GLM-5.2 targets the critical need for consistency in long-horizon, complex reasoning tasks.
Key Stories:
ServiceNow explores the vulnerability of research agents to data leakage through 'MosaicLeaks'. The study examines how agentic workflows can inadvertently expose sensitive information when aggregating data from multiple sources.
The Model Context Protocol now supports Enterprise-Managed Authorization, allowing organizations to centrally provision server access via identity providers. This enables zero-touch OAuth, connecting users to servers on first login without manual per-app authentication.
OpenAI reasoning models are being used to help physicians identify rare genetic diseases in children. In initial cases, the model helped secure 18 new diagnoses for previously unsolved medical mysteries.
Anthropic detailed seven primary methods for controlling Claude Code's behavior, including CLAUDE.md files, skills, and subagents. This framework allows developers to manage the authority and context cost of instructions for more predictable agentic behavior.
Hugging Face explores benchmarking open-source models on custom toolsets to evaluate their agentic capabilities. The post provides insights into how to measure a model's ability to effectively use specific developer tools.
Anthropic introduces organization-wide provisioning for MCP connectors via identity providers like Okta. This allows admins to centrally manage authorization and automate connector access for all users upon login.
Hugging Face analyzes alternatives to Low-Rank Adaptation (LoRA) for efficient model fine-tuning. The piece evaluates whether newer PEFT techniques can outperform LoRA in terms of performance and resource efficiency.
Claude Code now features live, interactive Artifacts that allow developers to preview in-progress work directly from the session context. These artifacts are shareable with teams, bridging the gap between CLI coding and visual feedback.
Amazon introduces Strands Agents and LeRobot, streamlining the pipeline from the Hugging Face Hub to physical robot hardware. This framework simplifies how developers deploy learned behaviors and agentic logic to real-world robotic systems.
Z.ai releases GLM-5.2, a model specifically optimized for long-horizon tasks and complex reasoning. The model aims to improve consistency and performance in extended workflows where maintaining long-term context is critical.
Anthropic introduces Workload Identity Federation (WIF), replacing static API keys with short-lived, scoped credentials. This improves security for enterprise deployments by integrating with OIDC-compliant identity providers.
Hugging Face launches Agentic Resource Discovery, enabling AI agents to search and find relevant resources more effectively. This expands the ecosystem's ability to support autonomous tool and data discovery.
Claude Design now integrates with design systems and Claude Code, enabling direct canvas editing and better brand consistency. It's accessible via the desktop app sidebar or claude.ai/design.
OpenAI has introduced Deployment Simulation to predict model behavior using real conversation data before public release. This methodology enhances safety evaluations and improves the accuracy of pre-deployment risk assessments.
Anthropic showcases winning projects from the Claude Code hackathon, demonstrating agentic applications in medical training, electronics repair, and factory maintenance. Highlights the practical utility of Claude Code in specialized technical domains.