Anthropic just gave its Claude Managed Agents platform two features that enterprise developers have been waiting for: scheduled deployments via cron schedules and secure environment variable storage through credential vaults. The update, which dropped around June 10, effectively turns these AI agents from on-demand assistants into autonomous workers that can execute tasks without anyone pressing a button.

The scheduled deployments feature uses cron schedules, a time-based job scheduling system that’s been the backbone of server automation for decades. This means use cases like automated research, continuous internal operations, and recurring data analysis can now run without manual triggering.

The second addition, environment variables managed through credential vaults, addresses a security problem that’s quietly plagued AI agent deployments. When agents need to access external tools, databases, or APIs, they need credentials. Anthropic’s approach isolates real credentials from the model environment entirely. The agent gets access to the tools it needs without ever directly handling the API keys or secrets that unlock those tools.

Claude Managed Agents launched in public beta on April 8, 2026, offering managed cloud environments and built-in tools designed to let developers focus on task design rather than infrastructure headaches. Since that launch roughly two months ago, the platform has been building out integrations. Cloudflare is among the platforms that have developed integrations leveraging Claude Managed Agents capabilities.

The credential vault feature is particularly relevant for any workflow touching financial systems. API keys for exchanges, blockchain RPC endpoints, and proprietary data feeds are all sensitive credentials that need careful handling. A system that keeps those secrets away from the model layer reduces the attack surface considerably.

The gap between public beta launch and production-grade features like scheduling and secure credential management was about two months. That’s fast. It suggests Anthropic is hearing loud and clear from enterprise customers that these capabilities aren’t nice-to-haves, they’re prerequisites for deployment.