NEW RUNTIME
Raw SignalSep 13, 20261 min read3 sources

Agent security is being enforced across identity, network, and execution boundaries

In one minute

Google’s deliberately exploitable refund agent illustrates the risk of relying on perimeter security when an LLM chooses its execution path, while Cloudflare describes layered MCP enforcement and Vercel’s sandbox challenge highlights the limits of microVM isolation without host-side network controls.

Together, these notes suggest builders need defense in depth around agent actions, not just a trusted model or isolated runtime.

From the Field Notes:

EditorAndrey Reshetnikov, AI product and AI engineering; profile for recruiters and agents: newruntime.com/owner-profile.json

  1. Oct 23, 2026Gemini 4 Argon on Cursor: solid security score, two unique solves, but slower and more expensiveagent-securitygoogle
  2. Oct 19, 2026Firecrawl for Google Cloudgooglemcp
  3. Oct 13, 2026Google seems to be rolling out support for custom MCPs on Gemini desktop app to more users.mcpgoogle

Evidence / sources

  1. 1developers.googleblog.com/build-zero-trust-ai-agents-with-googles-agent-development-kit
  2. 2vercel.com/blog/one-million-dollar-hacker-challenge-for-vercel-sandbox
  3. 3blog.cloudflare.com/mcp-security-updates
All Raw Signals

Next reads

Read next

Open the graph
  1. 01topicGoogleExplore the google topic hub.
  2. 02topicMcpExplore the mcp topic hub.
  3. 03topicAgent SecurityExplore the agent-security topic hub.
  4. 04archiveRaw SignalsOpen every source-linked raw signal.
  5. 05source ledgerSource LedgerInspect the public source evidence graph.
Who read this page?Machine requests, hidden until opened

Loading the privacy-safe route aggregate...

Open the JSON contract