MCP
30 days · UTC
Synchronizing with global intelligence nodes...
Harness rolls out Autonomous Worker Agents with governance, context, and a forkable marketplace
Harness introduced Autonomous Worker Agents for CI/CD with built-in governance, context, and an agent marketplace. Per [DevOps.com](https://devops.co...
Okta brings AI agent governance inside FedRAMP; identity-first agents meet enterprise reality
Okta moved AI agent governance inside FedRAMP boundaries, signaling identity-first agents are getting enterprise-grade controls. Okta says it’s the f...
Claude Code tightens MCP tool matching; ecosystem patches auth and metrics edges
Anthropic’s Claude Code changed how hooks match hyphenated MCP tool names and shipped a raft of reliability fixes. The latest Claude Code release [v2...
AWS Labs open-sources an agentic LLM evaluation system with multi-judge scoring
AWS Labs released an open-source, agent-guided LLM evaluation system that automates dataset creation, multi-judge scoring, and reporting. The new [AW...
SonarQube’s MCP server lands for Claude Code; 2.1.195 fixes risky tool matching
SonarQube now publishes an MCP server and generator for Claude Code, and Claude Code 2.1.195 tightens tool matching and agent stability. Sonar publis...
BYOK goes first-class: VS Code offline + Copilot multi-provider and request hooks
Microsoft and GitHub quietly made BYOK and offline LLMs first-class across VS Code, Copilot, and the Copilot SDK. VS Code 1.122’s new air‑gapped mode...
Codex v0.142.0 brings real agent governance: budgets, delegation gates, and allowlisted live search
OpenAI Codex added practical guardrails for coding agents and hardened its MCP/exec stack in v0.142.0. The latest [Codex release](https://github.com/...
Public Sentry DSN can hijack MCP agents in Claude Code, Cursor, and Codex
A new report shows a public Sentry DSN can hijack MCP-enabled agents in Claude Code, Cursor, and Codex. The New Stack details an agentjacking path wh...
Zep Graphiti shows a practical path to real-time agent memory—and a nudge toward portable skills
Zep’s Graphiti demonstrates real-time agent memory by combining knowledge graphs with vector-speed retrieval. This hands-on walkthrough builds a live...
Copilot SDK adds session memory and lazy tool loading
GitHub Copilot SDK now supports persistent session memory and lazy tool loading that leans on model tool search. In [v1.0.2](https://github.com/githu...
Claude Code v2.1.181: in-session /config, reliability fixes, and safer writes
Anthropic shipped Claude Code v2.1.181 with mid-session configuration and fixes for data-loss and flaky startups. The new release adds a simple /conf...
MCP turned scattered agent tools into a shared server with cleaner human gates
A real-world team replaced per-agent tool definitions with an MCP server to standardize discovery, versioning, and human approval paths. In this case...
OpenAI’s Responses API quietly rewrites how you build long-running agents
OpenAI’s Responses API moves agent workflows to server-side sessions and webhooks, cutting glue code and changing how you run long tasks. A developer...
MCP is making databases first-class tools for LLM agents
Vendors are shipping official MCP servers for major databases, making agentic, schema-aware DB access viable without custom glue. An [InfoWorld round...
Expose your catalog as an MCP tool or assistants won’t see your products
ChatGPT and other assistants increasingly shop via tools that query real product catalogs, not by scraping your site. A detailed walkthrough shows ho...
Agents got safer: Antigravity patches skill security; WindsurfAPI fixes WebFetch completion semantics
Antigravity hardened its popular skill library against path traversal and unsafe plugin behaviors, and related agent stacks tightened up tool-call han...
Windsurf becomes Devin Desktop, with smarter local context and honest rate‑limit signaling
Windsurf rebranded to Devin Desktop and shipped updates that change agent context handling and how clients back off under rate limits. Per the [chang...
OpenAI sets 2026 shutdown for Prompt Objects; start migrating and make agents portable
OpenAI will shut down Prompt Objects in its API on November 30, 2026, which means migration work for any agents depending on them. OpenAI confirmed t...
Agentic CLIs harden up: open, permissioned, and local-first
Agentic coding CLIs are quietly shifting toward safer, open, local-first workflows—and the ecosystem just shipped changes that make this concrete. A ...
MCP is moving into production fast — and Flowise’s 9.9 RCE shows the security cost of stdio agents
A 9.9‑severity RCE in Flowise’s MCP stdio path spotlights how agent tooling can become an OS‑level attack surface overnight. Researchers detailed a p...
GitHub Copilot switched to usage-priced credits; teams are seeing real cost spikes
GitHub Copilot moved from request-based billing to usage-priced credits, and heavy users are burning through quotas fast. Per GitHub’s rollout, month...
Copilot CLI tightens tool-call safety; GitHub app unblocks agent permission flows
GitHub Copilot CLI changed how tool calls are gated, and the GitHub app fixed permission dialogs that could stall agent runs. In the Copilot CLI pre-...
Snowflake is buying Natoma to put guardrails on MCP-connected AI agents
Snowflake is acquiring Natoma to bring identity, policy, and audit controls to MCP-connected AI agents across enterprise systems. Snowflake plans to ...