vibehacker
News
Google Blog ·

Gemini 4 Argon opens to developers via the paid Gemini API, AI Studio, and Antigravity

Two weeks after a cyber-defenders-only debut, Google's frontier Gemini 4 Argon is now available on the Gemini API paid tier, pitched at long-horizon work like code migrations and real-world software engineering. Google AI Ultra subscribers can also try it in AI Studio and the agent-first Antigravity platform, with Android Studio support announced alongside.

More news

View all

Azure MCP Server 3.0 rejects unknown parameters and creates Azure resources secure by default

Azure MCP Server 3.0.0 (Oct 12) now fails tool calls that pass dropped options like subscription or retry , creates Storage accounts with Shared Key off and Event Hubs with local auth off, and speaks the stateless MCP 2026 07 28 protocol. Pin @azure/[email protected] and clean up saved agent prompts before upgrading; the MySQL/PostgreSQL query tools are now marked destructive, so use read only if the agent should only look…

Start Debugging

GitLab Duo Agent Platform adds goal-driven orchestration across review, tests, and approvals

GitLab shipped agent orchestration in Duo Agent Platform: goal driven flows ( /goal in Duo CLI, headless, or Agentic Chat) run end to end through review, tests, security scans, and approvals under your identity and policy; Custom Flows are GA and can start from mentions, assignments, or pipelines. Needs GitLab 19.3+ and Duo CLI 9.17+ on Premium/Ultimate; GitLab for Slack and the MCP server are due GA later this month…

GitLab Blog

Vercel Labs’ fx: Zig coding agent that lists which MCP features it skips

Kin Lane covers Vercel Labs’ experimental Apache 2.0 fx agent (Zig, 6MB, v0.0.13 on github.com/vercel labs/fx ): it documents MCP transports and supported features, then states what it skips—Tasks, Apps, Skills over MCP, Client Credentials, Enterprise Managed Authorization, registry install, and acting as an MCP server. Default mode auto runs routine local work but routes every MCP tool call through a review model first…

API Evangelist

ChatGPT browser Annotations go extensible, paired with WebMCP site tools

Sites can customize selection targets, metadata, and up to 12 preview controls via document.oai.annotation in ChatGPT’s desktop built in browser, and register WebMCP site tools with document.modelContext.registerTool so Work/Codex get the selected context and call the page’s own actions. Docs target DevDay 2026+ desktop; Enterprise needs admin policy, Edu has no Site tools, and iframe or declarative HTML tools aren’t supported…

Spotted something we missed? Start a thread.