# Claude Code v2.1.330 adds Enterprise Admin Console sign-in and self-hosted runner options

published: 2026-11-20T23:05:38.311Z
topic: frameworks
source: https://github.com/anthropics/claude-code/releases/tag/v2.1.330
url: https://aiagentnews.io/stories/claude-code-v2-1-330-adds-enterprise-admin-console-sign-in-and-self-hosted

Anthropic's Claude Code v2.1.330 adds sign-in to organizations managed in the Enterprise Admin Console and a built-in chart mod called easel. The release also includes self-hosted runner options, new metrics, and fixes for freezes, lost settings and worktree deletion.

After running claude auth login, Claude Code now applies an organization's settings from the Enterprise Admin Console and uses that organization's cloud provider or gateway. A built-in mod called easel draws charts when asked and under MCP metrics results, on Pro, Max and, where available, Team and Enterprise plans on macOS and Linux. Automatic charts can be switched off in /config.

The self-hosted runner gains a --start-dir option, also set through SELF_HOSTED_RUNNER_START_DIR, so Claude begins work in a folder of a repository while settings and hooks load as usual. Two new /metrics values report seconds from a work order to registration and to the first session. Plugin install --json results now include commandPrinted and commandFailed fields.

Fixes cover a freeze of several minutes after compacting a very long conversation, ExitWorktree deleting a worktree that another session was using, and user settings changes being lost when several processes saved at once. Recurring scheduled tasks and /loop now run up to 30 days instead of 7 while their session runs, though tasks created before this release still end at 7 days.

## Why it matters if you build agents

Teams that use the Enterprise Admin Console can sign in once and have Claude Code pick up the organization's provider or gateway settings. Operators of self-hosted runners can start sessions in a chosen subfolder and track how long orders take to reach registration. Agents that rely on scheduled tasks now have a 30-day window, so check how long tasks you create will run and what notice they give when they end.
