Claude, the essentials — edition of October 25, 2026
Agent Skills moves to a neutral foundation as Claude reaches Google Cloud projects
Anthropic is handing the Agent Skills standard to the Agentic AI Foundation. Claude is also now available as an Anthropic-operated API inside Google Cloud projects, and Claude Code shipped several sandbox and compatibility fixes.
- Anthropic is donating Agent Skills (the name, logo, GitHub organization and agentskills.io domain) to the Agentic AI Foundation under the Linux Foundation. It reports 1 million published skills and nearly 280 million installs.
- Claude Models on Google Cloud (Anthropic managed) is the Claude API, operated by Anthropic and reached through a customer's Google Cloud project. Google Cloud IAM controls access and Google Cloud Marketplace handles billing. Fast mode is available there for Claude Opus 5.5.
- Claude Code 2.1.309 fixes three sandbox gaps. Version 2.1.310 extends the Monitor tool to Bedrock, Vertex and Foundry and adds a retry for gateways that reject certain request fields.
- The status page shows two short incidents on October 23–24, both resolved, and a broader elevated-errors incident across five models that was still marked as in progress.
- Anthropic opened a call for life sciences guest investigators. It also published two pieces whose titles concern an alignment assessment of a UK AISI cybertesting incident and formal software verification.
Agent Skills gets a vendor-neutral home
Anthropic announced on October 22 that it is donating Agent Skills, the open standard for giving agents specialized knowledge and workflows, to the Agentic AI Foundation (AAIF), part of the Linux Foundation. The donation covers the Agent Skills name and logo, the GitHub organization and the agentskills.io domain. Anthropic says it co-founded AAIF last year with Block and OpenAI, and that it had already donated the Model Context Protocol to the foundation. It says the aim is to keep the standard open, community-driven and vendor-neutral, and that people can use or contribute to it without joining AAIF.
A skill, as Anthropic describes it, is a folder containing a SKILL.md file that teaches an agent a task or domain, plus any scripts or reference material it needs. Anthropic says the idea came out of its work on Claude Code, which handled general knowledge work well but often lacked the procedural knowledge real business tasks require. It reports that more than 70 agent platforms, including ChatGPT, Cursor, GitHub Copilot and VS Code, have integrated the standard since it was published in December 2025. It also reports 1 million skills published and nearly 280 million installed. Going forward, engineers from Anthropic, Google, OpenAI and other companies will maintain the standard as individuals, with David Soria Parra and Mahesh Murag as Lead Maintainers. The proposed roadmap covers versioned releases, a formal proposal process and security guidance for people who write and install skills.
Sources: Donating Agent Skills to the Agentic AI Foundation — Anthropic
Platform: a new Google Cloud route, and a busy status page
Anthropic's API release notes for October 22 announce Claude Models on Google Cloud (Anthropic managed). Anthropic describes it as the Claude API, operated by Anthropic and reached through the customer's own Google Cloud project. Access is governed by Google Cloud IAM and billing runs through Google Cloud Marketplace. The same notes say fast mode is available for Claude Opus 5.5 on this offering.
The status page recorded several incidents over the same period. Two are resolved and short. Errors on Claude Opus 5.5 lasted from 17:23 to 17:34 PT on October 23 (00:23 to 00:34 UTC on October 24), about eleven minutes. Errors on Claude Fable 5 and Claude Fable 5.1 lasted from 22:00 to 22:35 PT the same evening (05:00 to 05:35 UTC on October 24), about thirty-five minutes. A third incident, titled as elevated errors for Claude Sonnet 5.5 and other models, was still marked as in progress. It names Claude Mythos 5.1, Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5. The entry says only that Anthropic is investigating and will update as soon as possible; it gives no cause and no resolution.
Sources: Claude Models on Google Cloud (Anthropic managed) – API release notes, October 22, 2026 — Anthropic (release notes) · Elevated errors for Claude Sonnet 5.5 and other models — Claude Status · Elevated errors for Claude Fable 5 and Claude Fable 5.1 — Claude Status · Elevated errors for Claude Opus 5.5 — Claude Status
Claude Code: a tighter sandbox and wider compatibility
Version 2.1.309 of Claude Code is mostly about the sandbox. It fixes sandboxed Bash commands on macOS and Linux running the shell's startup files outside the sandbox before each command. It fixes macOS denyWrite glob patterns not applying at a path's real location when the pattern's folder is reached through a symlink. It also fixes sandboxed commands and auto-accepted edits being able to modify the MCP config file passed with --mcp-config. Each of these concerns a boundary users may have assumed was already holding.
Version 2.1.310 turns to deployment environments. The Monitor tool and plugin monitors are now available on Bedrock, Vertex and Foundry, and when telemetry or nonessential traffic is disabled. Long foreground sleeps are now blocked in those configurations too. Claude Code also now retries without the effort-2025-11-24 beta value or mid-conversation system messages when a gateway or compatible server rejects them, which previously made requests fail. The release also fixes background cleanup deleting some recent files, such as pasted images, when cleanupPeriodDays is set to a very large value. Earlier, version 2.1.307 added /claude-api api-version, which moves Claude API code to API version 2026-10-21.
Sources: Claude Code 2.1.310 — Claude Code (GitHub) · Claude Code 2.1.309 — Claude Code (GitHub) · Claude Code 2.1.307 — Claude Code (GitHub)
Research outreach and two publications
On October 22 Anthropic opened a call for guest investigators in the life sciences. It invites principal investigators and their groups at academic institutions, nonprofits or government labs to propose projects in areas such as causal human biology, molecular biology, medicinal chemistry, drug discovery and clinical science. Selected teams would work directly with Anthropic researchers for 8 to 12 weeks. Anthropic would run its models and harnesses at no cost to them, and the outputs would be published and the data and code open-sourced. Guest investigators keep the rights to their data and to any independent follow-on research. Applicants are asked not to supply their own data and to include only publicly available information. Applications stay open for about 14 days, and Anthropic aims to notify selected participants about 14 days after they close.
Two further Anthropic publications reached us with only their titles, so we describe nothing beyond them. One is an alignment assessment of the UK AISI cybertesting incident. The other is headlined as arguing that LLMs are making formal software verification feasible at a much larger scale. The material provided includes no summary, findings or context for either. Readers who want the substance, including any cause or circumstances Anthropic gives for the incident, should go to the original pieces.
Sources: Open call for guest investigators in the life sciences — Anthropic · An alignment assessment of the UK AISI cybertesting incident — Anthropic · LLMs are making formal software verification feasible at a much larger scale — 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.