Cybersecurity
A secure workspace where your team runs Claude on client files instead of pasting them into a public chatbot. Includes the AI use policy and a build guide for partners, CFOs, and CISOs.
Follow the setup below in your own workspace. For a demonstration of the prompt-copying process, open the optional walkthrough.
Watch the setup walkthroughNever used Claude? It is free and takes 30 seconds to open. Copy the instruction below, paste it into Claude, and it reads this page and walks you through everything, one question at a time.
One tap. It changes how much Claude explains, and how slowly it goes. You can change it any time.
A short instruction plus a link to this page lands on your clipboard. First copy asks for your email once. That unlocks every button across the whole library.
Free account, no card, 30 seconds. This tab stays open so you can come back.
Open claude.aiClaude reads this page, asks which computer you are on, then guides you step by step until it works. If anything errors, tell Claude what you see, and it fixes it with you.
Install Private Claude Setup for Client Data on my computer. Walk me through it. Repo: https://platform.claude.com/docs/en/managed-agents/self-hosted-sandboxes What it does: A secure workspace where your team runs Claude on client files instead of pasting them into a public chatbot. Includes the AI use policy and a build guide for partners, CFOs, and CISOs. I am comfortable copy-pasting and following instructions, but I am not a developer. Rules: - Plain English. Define jargon the first time it appears (repo, env var, port, dependency). - One step at a time. Exact command in a code block. Tell me which app to paste it into (Terminal on Mac, PowerShell on Windows). - One sentence per command explaining what it does and what success looks like. - After each command, wait. I will tell you the output before you move on. - If a tool is missing (git, node, docker, python), give me the one-line install for my OS first. - If something errors, diagnose before the next step. Do not skip. First message: ask only "What is your operating system, macOS, Windows, or Linux?" Then start step 1. Reference steps from the public guide (adapt to my OS, do not just paste them at me): 1. The decision first: give your team ONE private workspace your firm controls, instead of 12 free accounts on personal phones. That single move is what stops the unlogged leak. 2. What to tell your team: client files stay in the firm workspace, never a free chatbot. Keep a record of what the AI touched. That is the confidentiality answer you give a client before they ask. 3. For whoever implements it (the technical guide, button below): self-hosted sandboxes keep tool execution and your files on your own infrastructure; MCP tunnels let agents reach internal systems over one outbound encrypted connection, no public endpoint. 4. Lock it down: scope access with least privilege, container-per-session for sensitive data, never put the org API key on the worker host, keep your own audit trail. 5. The honest ceiling: the model inference still runs on Anthropic (your files and tool execution stay in your environment, and Zero Data Retention stops prompt storage, but the chat is not fully on-prem). Both features are preview/beta. A vendor cert is not your compliance. Test before crown-jewel use. Stop when the app opens and I confirm it works.
The code is public and free. The setup instruction above installs and wires it for you. You never need to open this link.
Browser-based. No signup. Drop in your numbers and see the trade in real time. Opens in a new tab so the prompts stay where you left them.
Regulated teams banned AI because nobody could answer where the data lives. Claude's self-hosted sandboxes and MCP tunnels run agent tool execution on your own infrastructure over a single outbound encrypted connection — the answer that unstalls the pilot.
The setup rail and every prompt above are free and stay free. The cost is your time, and the risk of wiring it wrong on live data.
Back to the promptsWe wire it into your stack: gateway placement next to your MCP servers, egress policy on the sandbox, RBAC scoping on every tunneled system, container isolation per session, and an audit trail your auditor accepts. The pilot ships inside your boundary without a misconfigured tunnel becoming the incident.
Book a build callSelf-hosted sandboxes are public beta and MCP tunnels research preview as of writing; verify status and Zero Data Retention / HIPAA BAA eligibility against current Anthropic docs. Anthropic holds SOC 2 Type II, ISO 27001, ISO 42001, HIPAA; your own deployment compliance remains your responsibility.
Want this wired into your stack instead of running it yourself? That is our AI implementation services service.
Newsletter
One short note when a new build ships: what it does, who it is for, and how to run it in your own Claude.
Free. One email a week at most. Unsubscribe in one click.
More like this
754 cybersecurity skills for AI coding agents like Claude Code and Cursor, mapped to MITRE ATT&CK and NIST, plus a calculator that finds critical CVEs in your stack.
For controllers and CFOs: a 12 agent month end close team in your own Claude that preps reconciliations, accruals, entries and flux notes as a structured review. You approve every ledger post.
For real estate fund managers and family offices: 13 Claude agents to prep deal screens, lease abstracts, capital calls and LP letters as a structured review. You approve anything that moves money.
Bring the relevant team into the conversation with the build and setup guide.
Private Claude Setup for Client Data is a cybersecurity build in the consultance.ai AI Build Library. A secure workspace where your team runs Claude on client files instead of pasting them into a public chatbot. Includes the AI use policy and a build guide for partners, CFOs, and CISOs. It fits Managing partners, CFOs, GCs, and firm owners who are liable for client confidentiality, plus the CISOs, CTOs, and platform teams who implement it, in finance, legal, and accounting. Setup difficulty is Technical, with 5 plain-English steps.
A secure workspace where your team runs Claude on client files instead of pasting them into a public chatbot. Includes the AI use policy and a build guide for partners, CFOs, and CISOs.
It fits Managing partners, CFOs, GCs, and firm owners who are liable for client confidentiality, plus the CISOs, CTOs, and platform teams who implement it, in finance, legal, and accounting.
Technical to set up — one guided setup instruction covering 5 plain-English steps on the resource page.
We would wire it into your stack: gateway placement next to your MCP servers, egress policy on the sandbox, RBAC scoping on every tunneled system, container isolation per session, and an audit trail your auditor accepts.
Self-hosted sandboxes are public beta and MCP tunnels research preview as of writing; verify status and Zero Data Retention / HIPAA BAA eligibility against current Anthropic docs. Anthropic holds SOC 2 Type II, ISO 27001, ISO 42001, HIPAA; your own deployment compliance remains your responsibility.
Discuss how Private Claude Setup for Client Data fits your systems, controls, and team. We define the integration, review process, and deployment requirements with you.
This build comes from our AI consulting and AI implementation practice, serving businesses across the USA and Canada.