Anthropic said on 6 November that Claude now sets up a repository to run in Claude Code cloud sessions by itself on its Pro, Max and Ultra plans. “Claude figures out what dependencies, network access, and API keys it needs, then saves the environment for next time”, its ClaudeDevs account posted (ClaudeDevs).
The post added: “API keys are saved securely, so Claude never sees them.” The change covers three plans, by our count; Anthropic’s post does not mention Team or Enterprise.
Business Pill · AI Agents
An evergreen Business Engineer explainer on AI agents. It teaches the general idea only and does not discuss Anthropic or Claude Code.
The key insight: Anthropic says Claude Code cloud sessions on Pro, Max and Ultra now work out the dependencies, network access and API keys a repository needs and save that environment for next time; its docs say stored keys are added by an agent proxy outside the session’s VM.
Anthropic’s documentation describes a Default cloud environment that “carries no configuration of its own”: it “defines no environment variables or setup script”, and users configure an environment when Claude “needs dependencies installed before it starts working” (Claude Code docs). The 6 November post says Claude now works out those needs and saves the result.
Anthropic says Claude now works out the dependencies, network access and API keys a repository needs in a Claude Code cloud session and saves the environment for next time; the docs say stored keys are added by an agent proxy outside the session’s VM. Data: @ClaudeDevs, 6 Nov 2026; Claude Code docs.
What Was Posted on X
We’ve made it easier to get your repo running in Claude Code cloud sessions on Pro, Max, and Ultra.
Claude figures out what dependencies, network access, and API keys it needs, then saves the environment for next time. API keys are saved securely, so Claude never sees them.
Anthropic’s ClaudeDevs account announced the change at 19:58 UTC on 6 November: Claude figures out the dependencies, network access and API keys a repo needs in Claude Code cloud sessions on Pro, Max and Ultra, and saves the environment for next time.
Claude on Pro, Max and Ultra plans now can figure out your project dependencies and required API keys on its own in order to run it in the Claude Code cloud session.
— ๐จ AI News | TestingCatalog (@testingcatalog) Nov 06 23:14
TestingCatalog posted at 23:14 UTC on 6 November, 3 hours 16 minutes after Anthropic’s post, that Claude on Pro, Max and Ultra plans can figure out project dependencies and required API keys on its own in Claude Code cloud sessions.
How the Keys Stay Out of Claude’s Reach
The documentation describes the mechanism for stored keys, which it calls network secrets: “Anthropic’s agent proxy adds the key to requests for the hosts you list, after each request leaves the session’s VM, so the key itself stays outside the VM.” The docs say network secrets are available on Pro, Max, Ultra, Team and Enterprise plans.
Who Spotted It
TestingCatalog flagged the change on X at 23:14 UTC on 6 November, 3 hours 16 minutes after Anthropic’s post, by our count.
Our Numbers
Three plans named in Anthropic’s post (Pro, Max and Ultra), three things Claude now works out by itself (dependencies, network access and API keys), and a 3-hour-16-minute gap between Anthropic’s post and the relay. All counts are ours.
Connecting the Coverage
Claude Code has had a run of changes this week: an email check that arrived 108 days after the first report (our piece) and subagent status updates that now cost about 99% less (our piece). Why the product matters to Anthropic’s revenue is in our earlier analysis (our piece).
The Structural Read
As we read Anthropic’s post, setup that a user configured by hand in a cloud environment is now something Claude works out and saves.
As we read the documentation, the key-handling claim rests on network secrets: Anthropic’s agent proxy adds the key to requests after each request leaves the session’s VM.
As we read the post, the change is scoped to three plans, by our count, and the post does not mention Team or Enterprise.
Anthropic (@ClaudeDevs), 6 November 2026
“API keys are saved securely, so Claude never sees them.”
Three Implications
FOR PRO, MAX AND ULTRA USERS Anthropic says Claude saves the environment it works out for next time.
FOR API KEYS Anthropic’s documentation says network secrets stay outside the session’s VM and are added by its agent proxy.
FOR TEAM AND ENTERPRISE Anthropic’s post does not mention these plans, so whether the automatic setup reaches them is not established.
The analysis says: “The surface, not the model, governs autonomy.”
As we read it, a cloud session that works out its own dependencies, network access and keys is a change to the surface Claude Code runs on, not to the model. The analysis does not discuss this feature.
What Is Not Established
Anthropic’s post does not say whether the automatic setup is coming to Team and Enterprise plans, how Claude chooses the network access it requests, or whether users approve each saved setting. We have not tested the feature.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
Anthropic says Claude Code cloud sessions on Pro, Max and Ultra now work out the dependencies, network access and API keys a repository needs and save that setup for next time, with keys stored so Claude never sees them. The documentation says stored keys are added by a proxy outside the session’s VM.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. The capability and plans are from Anthropic’s ClaudeDevs post on X of 6 November 2026; the Default environment and network-secrets mechanism are from Anthropic’s Claude Code documentation. All counts are ours. We have not tested the feature. This is not investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
