531 posts tagged with "AI agents"
AI agents and autonomous systems
Railway Cloud Agents Promote Your Dev VM Straight to Prod: What Skipping the Build Actually Costs
Railway's Cloud Agents Beta promotes the agent's lived-in VM straight to production via a VM daemon, skipping git push and CI as too slow. A five-row ledger of what the build step was quietly doing — drift, secrets, approvals, rollback, provenance — and how to keep the fast loop without shipping a mutable machine.
Your Sandbox Is Not a Home: Where Vibe-Coded Apps Actually Live After Day One
Every 2026 guide to shipping vibe-coded apps ends at 'deploy behind a sandbox' — without saying where the app lives long-term. The missing fifth stage: a permanent git-backed home with a redeploy loop, the sandbox-vs-home checklist, and the cost crossover between execution meters and flat hardware.
SmolVM vs. Firecracker: Which MicroVM Should Actually Isolate Your AI Agent's Code?
SmolVM's sub-200ms boot challenges Firecracker's two-year default. A measured comparison of boot time, memory density, and what actually isolates untrusted agent code from the host kernel across SmolVM, Firecracker, OpenSandbox, and Docker Sandboxes.
Run a Claude Fable Security Scan with Bex Security
Run a Claude Fable security scan with Claude Code and Bex Security. The exact command, the model row that actually runs, and the sandbox policy Bex enforces around Claude.
Defang's `/deploy` Skill Reveals the Three-Layer Stack Agent-Native PaaS Needs
Defang's deploy skill shows how agent runbooks, MCP actions, Docker Compose, and open Pulumi providers form an auditable agent-native PaaS stack.
Feltsense's $5.1M Bet on AI Founders: What Agents Can Run—and Where Humans Still Sign
A concrete map of what Feltsense-style AI founders can operate today, where humans remain legally required, and which metrics would prove the $5.1 million bet works.
Before AI Agents Operate Your Platform: The Catalog Hygiene Contract Your IDP Needs
Use a field-level catalog contract, a worked deploy example, and staged authority controls to let AI agents safely operate a self-hosted developer platform.
Anthropic Paused Its Claude Agent SDK Meter. Your Agent-Ops Cost Model Still Needs It.
Anthropic paused its proposed Agent SDK credit split, but deploy-from-chat platforms still need a concrete model for attributing and budgeting unattended agent work.
Gateway API 1.6 Gives Agent Sandboxes a Named Egress Target—Not an Egress Policy
Gateway API 1.6's experimental XBackend can make an AI API destination reviewable. Here is the complete control stack required to turn that declaration into safe sandbox egress.