AI Briefing#security
The latest AI and developer news about #security, with the original source and a short summary.
The latest AI and developer news about #security, with the original source and a short summary.
Vercel has officially launched a security dashboard that provides a unified view of security posture for accounts and projects across all plans.
This article presents an evaluation framework for verifying the real-world performance and safety of LLMs before production deployment.
Vercel announced that it has completed protective measures for the Next.js security vulnerabilities disclosed in August 2026, requiring no action from customers.
Vercel has launched the Run SDK to safely execute untrusted code.
Vercel Connect addresses agent security issues with short-lived, runtime-based tokens.
It has been raised that LLMs can control host machines by exploiting code execution vulnerabilities in inference engines.
CopilotKit has released OpenBot, a self-hosted platform featuring agent-specific isolated environments and tool-call auditing capabilities.
Vercel has strengthened security management by categorizing environment variables into Config and Secret types.
A remote code execution (RCE) vulnerability has been discovered in the Unitree Go2 robot, raising the risk that an entire robot fleet could be hacked.
Cloudflare has introduced a feature that allows selective permission of specific scopes in OAuth clients.
When searching for Claude Code installation, malicious artifacts ranked high on Google, leading to the installation of a macOS information-stealing tool.
v0 blocks generated code from accessing user OAuth tokens by routing requests through a proxy outside the sandbox.
Vercel is running a two-week public hacking challenge that pays up to $1 million to researchers who discover sandbox isolation bypass vulnerabilities.
Vercel added a compliance documents section to team settings, enabling direct downloads from the dashboard.
Securely sign JWTs in Vercel Functions without exposing private keys using Vercel KMS.
Vercel CDN has started supporting ECH, which enhances security by encrypting the SNI in the TLS handshake.
senv, an OS-level Python sandbox tool that enables Claude Code to safely use uv workflows, has been released.
Cases have been discovered where Greptile inserts prompt injections into PR bodies to induce AI agents to autonomously modify code.
Claude Fable 5 achieved a 44.3% Python pass rate in JetBrains' private repository evaluations, surpassing Opus 4.8's 28.2%.
Anthropic's text watermark is hidden through statistical biases in word choice rather than at the byte level.
The update is currently in beta for Claude Enterprise customers and uses the existing Compliance Access Key.
DoorDash built an Agent Gateway to complement the operational limitations of MCP.
A Firestore misconfiguration on tl;dv exposed meeting metadata and some calls.
Anthropic has introduced inference hooks for Claude Enterprise, a new security feature available in beta that enables inline data loss prevention across chat, Claude Code, and Cowork by integrating with existing DLP servers.