FR
Live

Claude, the essentials — edition of November 18, 2026

Linked billing enters beta as Claude Code tightens safeguards and Meetings strains

By ·

Anthropic's API gained a beta that lets customers pay for Claude usage in third-party apps through their own Claude organization. Three Claude Code releases focused on safeguards and clearer messages, and Claude Meetings was still under mitigation after heavier-than-expected demand.

Key points

Claude Code: three releases that tighten the guardrails

The three Claude Code releases in the day's items, 2.1.325, 2.1.326 and 2.1.327, mostly concern what happens before an action takes effect and how clearly the tool explains itself. The one explicitly labelled a security fix is in 2.1.325, which corrects how zstd-compressed responses are decoded. Version 2.1.327 adds a confirmation before /install-github-app changes a repository. It also adds a form of claude mcp add-json that reads the server's JSON from stdin, so that tokens can stay off the command line. That form is not available inside a Claude Code session.

The other changes make the tool more explicit about its state. Version 2.1.326 warns at startup when an --allowedTools, --disallowedTools or --tools value contains an entry that does not look like a complete rule. It also explains why MCP servers stop loading when an administrator adds managed-mcp.json during a session, and it allows Code Mode to be turned on while agent teams are enabled. In 2.1.325, an error shown when an organization's policy blocks a request now says how to continue: run /rewind and change the message, or run /clear. Version 2.1.327 also adds a line when a resumed session's earlier messages cannot be loaded.

Sources: Claude Code 2.1.325 release notes — Claude Code · Claude Code 2.1.326 release notes — Claude Code · Claude Code 2.1.327 release notes — Claude Code

API: customers can now bring their own Claude organization to your app

Anthropic launched Linked billing in beta, enabled with the user-profiles-billing-2026-10-01 beta header. Its stated purpose is that a developer's customers can pay for the Claude usage in that developer's app through their own Claude organization. Once a customer links their organization, developers set billing_policy on each /v1/messages request. The excerpt we have continues to a further /v1 endpoint, and the rest of the note is not in the provided material.

A separate API release note, dated the day before, concerns provenance. Text generated by Claude Opus 4.6, Claude Opus 4.5, Claude Sonnet 4.5 and Claude Haiku 4.5 carries Anthropic's text watermark on supported platforms. The note points to the page How Claude marks AI-generated content for the list of those platforms. It then says what happens when a request to one of these models sets temperature, top_p or a similar sampling parameter. The excerpt is cut off before stating the consequence, so we do not report it.

Sources: API release notes, November 17, 2026: Linked billing in beta — Anthropic (release notes) · API release notes, November 16, 2026: text watermark — Anthropic (release notes)

Reliability: Meetings demand outruns capacity, two other incidents closed

The incident still listed as ongoing is Claude Meetings (beta). According to the status page, it saw higher than expected demand and Claude was unable to join many calls. Anthropic says it has put mitigations in place and is continuing to monitor and improve join reliability. Meetings is now re-enabled for all Enterprise customers, though Claude may take longer than usual to join. Adding Claude to new calls and Auto-join remain paused for Team customers.

Two other incidents are marked resolved. Sign in with Apple was failing for some users on claude.ai and the Claude iOS app. Anthropic said sign-in with Google and email were not affected. Separately, elevated errors for Claude Sonnet 5.5 and Claude Fable 5.2 affected claude.ai, the Claude API, Claude Code and Claude Cowork. Impact ran from 23:32 PT on November 15 (07:32 UTC on November 16) to 00:45 PT (08:45 UTC) on November 16.

Sources: Claude: Claude unable to join meetings — status.claude.com · Claude: Some users unable to sign in with Apple on claude.ai and the Claude iOS app — status.claude.com · Claude: Elevated errors for Claude Sonnet 5.5 and Claude Fable 5.2 — status.claude.com

Two Anthropic publications on disclosure and safety evaluation

Anthropic published two items that bear on its security and safety practice. One is titled Updating our coordinated vulnerability disclosure policy timelines. The other is titled Building strong safety evaluations.

The material provided for this edition contains only the titles of these pieces, not their contents. We therefore cannot say what the new disclosure timelines are or what the evaluation guidance recommends, and we do not speculate.

Sources: Updating our coordinated vulnerability disclosure policy timelines — Anthropic · Building strong safety evaluations — Anthropic

This edition is an original synthesis written by Claude from aggregated news — Anthropic's own sources first (release notes, status, newsroom, research, engineering), then the press, Hacker News, Reddit and GitHub, under the editorial supervision of Héra SASU. Every fact links to its article, publisher named. See the live feed →

Claude News is published by Héra SASU. Independent media, not affiliated with Anthropic.