Updates: #Agents.

[NEW] Building commerce agents with Claude · Anthropic released a blueprint for building commerce agents on Claude on September 2, 2026. It ships two agents defined once and deployable three ways — a shopping agent for customers and a merchant agent for staff, running on the Messages API, the Claude Agent SDK, or Claude Managed Agents — along with four runnable vertical examples covering retail, travel, telecom, and entertainment. A Claude Code plugin called commerce-builder scaffolds a custom agent for your own store. The code is available under Apache License 2.0 at anthropics/commerce-agents, and every example runs on fictional data with no live orders or charges.

[NEW] What's new in Claude Managed Agents — scheduled deployments and environment variables in vaults · Two capabilities land in public beta on the Claude Platform. Scheduled deployments give an agent a cron schedule — each time the schedule fires, the agent starts a new session and completes its task, with no scheduler for you to build or host. Environment variables in vaults store API keys tied to specific approved domains: the agent never sees your key because the sandbox only holds a placeholder, and the real key is attached at the network boundary, only on requests to domains you allow. Supports the Browserbase, KERNEL, Notion, Ramp, and Sentry CLIs plus any CLI using HTTP authentication.

[NEW] Dynamic workflows in Claude Code · Claude breaks a problem into subtasks and fans them out across tens to hundreds of subagents running in parallel, dynamically writing orchestration scripts that run in a single session and checking its work before anything reaches you. It targets broad work like codebase-wide bug hunts and security audits and large migrations spanning thousands of files, verifying each result before folding it into the final answer. Available as a research preview on Max, Team, and Enterprise plans.

[NEW] Claude Managed Agents — Self-hosted sandboxes and MCP tunnels · Claude Managed Agents now supports self-hosted sandboxes (public beta) and MCP tunnels (research preview). The agent orchestration loop stays on Anthropic infrastructure while tool execution runs in customer infrastructure or with managed sandbox providers — Cloudflare, Daytona, Modal, Vercel — and agents reach private-network MCP servers through a lightweight gateway with a single outbound connection, no inbound firewall rules required.

[NEW] Agent view in Claude Code · Agent view is a centralized interface inside the Claude Code CLI for managing multiple concurrent agent sessions. Now in Research Preview, it lets developers see which agents need input, are actively working, or have completed tasks — and scale up background work via peek, `/bg`, and `claude --bg`.

[NEW] What's new in Claude Managed Agents — Dreaming, Outcomes, Multiagent, Webhooks · A month after the April 8 beta launch, Managed Agents added two new features and promoted three from research preview to public beta. Dreaming (research preview) curates memory between sessions for self-improvement, and outcomes, multiagent orchestration, memory, and webhooks all ship as public beta.

[NEW] Claude Financial Services Agents — 10 templates, Moody's MCP, and 17 data connectors · Ten ready-to-run financial services agents — pitch builder, KYC screener, month-end closer, and more — ship as plugins in Claude Cowork and Claude Code, and as cookbooks for Claude Managed Agents. A Moody's MCP app brings proprietary credit ratings and data on 600M+ companies, and 17 data partners are integrated alongside.

[NEW] Claude Managed Agents · A new API alongside Messages that provides a pre-built agent harness and managed infrastructure — run Claude agents without building your own loop, sandbox, or tool-execution layer.