Code Mode in Claude Code
Claude Code adds an experimental opt-in Code Mode where Claude writes JavaScript to call its tools.
Loading launches from 4 labs
Topic
Systems that plan and take multi-step actions with tools
Claude Code adds an experimental opt-in Code Mode where Claude writes JavaScript to call its tools.
Self-hosted Claude Managed Agents allow you to run agents in your own cloud or on-premise environment. You host the agents, credentials, and session history while controlling where requests are sent.
A new `permissions.denyInsteadOfPrompting` setting automatically denies tool calls that would otherwise pause for permission prompts, letting the turn continue without interruption.
Claude Enterprise introduces security center, inference hooks, centralized management, and new access controls, with several features in beta for enterprise security teams.
Organizations can now create IP allowlists in the Claude Console to limit API requests to specific IPv4/IPv6 addresses or CIDR ranges, reducing risk from leaked keys.
Claude Code cloud sessions now display form cards when an MCP server requests input, allowing users to submit answers or decline rather than having requests canceled unseen.
Anthropic added guardrail.input_tagging to Claude apps gateway Bedrock upstreams. It wraps prompts in the input tags required by Amazon Bedrock guardrail prompt attacks filters.
The Claude apps gateway now supports oidc.user_id_claim, letting organizations identify users with an id_token claim other than sub, such as Microsoft Entra ID's oid.
Anthropic's Compliance API now includes a Security posture endpoint in beta for Claude Enterprise orgs, letting admins fetch current security posture results.