Sophos Central MCP

Sophos Central + AI - for ChatGPT, Claude, GitHub Copilot, Microsoft 365 Copilot, Gemini, and any agent that speaks MCP

Unofficial. Community-built Claude Code Skill and MCP server for the Sophos Central API. Not affiliated with, endorsed by, or sponsored by Sophos Limited.

_{▶ Watch the 30-second demo - demo data is simulated; every command shown exists in the real CLI.}

The Sophos Central partner, organization and tenant APIs, routed to the right tenant and region automatically, plus a cross-tenant triage queue, fleet tamper remediation, a multi-tenant SIEM collector and hostname-based containment. Works with the AI you already use - ChatGPT (Plus/Pro+, paused until this connector is reprinted; see mcp-install.md), Claude Desktop, Codex, Claude Code, Claude Cowork, and GitHub Copilot - plus Microsoft 365 Copilot / Copilot Studio and Google Gemini via the remote path. Free, open source, runs on your laptop. Built for MSP owners. No code required.

Works with your agent

Security note (until this connector is reprinted). Do not run HTTP mode (--transport http), not even on localhost. This build checks a token, but it predates the fix release's request size and time limits. The MCP tools can also make changes with your credentials. Use the default stdio transport. If you set up a tunnel for a remote agent, shut it down. A fix is on the way.

The six agents MSP owners actually use (self-serve, works today):

Your AI agent	How to install the Sophos Central skill
Claude Desktop	Run installer, then Settings > Extensions to register sophos-central-mcp (no JSON editing).
ChatGPT (paid plans)	Paused until this connector is reprinted: ChatGPT needs its HTTP mode, and in this build that mode checks a token but has no request size or time limits yet. See the security note in mcp-install.md.
Codex CLI	Paste the install prompt below.
Claude Code	Paste the install prompt below.
Claude Cowork	Paste the install prompt below.
GitHub Copilot (VS Code)	Run installer, add sophos-central-mcp to mcp.json under the servers key, then pick Agent mode.

ChatGPT needs a remote HTTPS server, which means this connector's HTTP mode, and that is paused until the connector is reprinted: in this build that mode checks a token but has no request size or time limits yet. See the security note in mcp-install.md.

Also for the Microsoft and Google stacks

Big install base, but an honest heads-up: these are the remote / enterprise path, not the local binary you just installed.

Agent	What it takes
Microsoft 365 Copilot / Copilot Studio	Paused until this connector is reprinted: Copilot Studio needs its HTTP mode, and in this build that mode checks a token but has no request size or time limits yet. See the security note in mcp-install.md.
Google Gemini	Gemini CLI is local - same as Claude Code. The Gemini app is remote and needs the same paused HTTP mode as ChatGPT. See mcp-install.md.

Skill-native agents (also covered): Hermes and OpenClaw read this skill's SKILL.md directly and speak MCP - see their install sections below. Also works with Cursor, Windsurf, Cline, Continue.dev, and Zed via MCP. Full per-tool wire-up: docs/which-agent.md.

Run more than one agent? Install across all 51+ supported agents in one command: npx skills add Servosity/msp-skills@latest (requires Node.js, then run the per-skill installer for the CLI/MCP binaries). See docs/which-agent.md.

Install in 60 seconds

Fastest for Claude Desktop - one-click .mcpb

Download Sophos Central MCP (.mcpb) - then open Claude Desktop > Settings > Extensions and select the file. One click, no JSON, no shell. (Browse every Sophos Central release on the releases page.)

Prefer the Claude Code plugin? Add the marketplace once, then install - works immediately, no directory listing required:

/plugin marketplace add Servosity/msp-skills
/plugin install sophos-central@msp-skills

Path A - paste one prompt into your AI agent (recommended)

Copy this into Claude Code, Codex CLI, or Claude Cowork:

Install the Sophos Central Skill and MCP server from Servosity/msp-skills in this agent workspace. If this workspace uses a POSIX shell (macOS, Linux, WSL, or Bash), run bash <(curl -fsSL https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/sophos-central/install.sh). If it uses Windows PowerShell, run iwr -useb https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/sophos-central/install.ps1 | iex. Then authenticate per the README and run sophos-central-cli --help to explore.

The same prompt works in any agent that can run shell.

Path B - run the installer yourself

Windows (PowerShell):

iwr -useb https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/sophos-central/install.ps1 | iex

macOS / Linux:

bash <(curl -fsSL https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/sophos-central/install.sh)

The installer drops both sophos-central-cli (the CLI) and sophos-central-mcp (the MCP server) into your user bin path. Claude Code, Codex, and Cowork discover the Skill via SKILL.md in this directory.

Verify:

sophos-central-cli --version

Upgrade to the latest version

The installer always fetches the current release - re-run it to upgrade:

macOS / Linux:

bash <(curl -fsSL https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/sophos-central/install.sh)

Windows (PowerShell):

iwr -useb https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/sophos-central/install.ps1 | iex

Claude Desktop .mcpb users: download the latest .mcpb (top of this section) and re-select it in Settings > Extensions. Claude Code plugin users: /plugin update sophos-central@msp-skills.

Add to Claude Desktop, GitHub Copilot, Gemini CLI, Microsoft 365 Copilot, or another MCP client

After the installer runs, see mcp-install.md and docs/which-agent.md for the per-agent wire-up - one section per agent, including the GitHub Copilot servers key and the remote Microsoft 365 Copilot / Copilot Studio path. Claude Desktop's Settings > Extensions panel is the simplest path; the MCP config block (for users who prefer editing JSON) is documented in mcp-install.md.

Install for Hermes

From the Hermes CLI:

hermes skills install servosity/msp-skills/skills/sophos-central --force

Inside a Hermes chat session:

/skills install servosity/msp-skills/skills/sophos-central --force

Hermes speaks MCP natively, so it can also use the sophos-central-mcp server directly - same install path, same env vars.

Install for OpenClaw

Tell your OpenClaw agent (copy this):

Install the sophos-central skill from https://github.com/servosity/msp-skills/tree/main/skills/sophos-central. The skill defines how its required CLI (sophos-central-cli) can be installed via the openclaw: frontmatter block.

OpenClaw isn't generally available yet; the frontmatter wiring is pre-shipped and will activate the moment OpenClaw launches.

Authenticate

Set the credentials the CLI needs (from your Sophos Central portal):

SOPHOS_CENTRAL_CLIENT_ID=<value> SOPHOS_CENTRAL_CLIENT_SECRET=<value> sophos-central-cli doctor

Create the credential in Sophos Central first: Partner Dashboard > Settings & Policies > API Credentials for a partner, or Global Settings > API Credentials Management for an organization or a single customer. doctor checks config, paths and API reachability, then mints a token with your credential and reports whether Sophos accepted it. A partner or organization credential picks one customer for tenant-scoped commands with --tenant <tenant-id> (or SOPHOS_CENTRAL_TENANT_ID); the fleet commands walk every customer on their own.

What this skill does

Question your MSP keeps asking	Command
What needs attention across all my customers today?	sophos-central-cli fleet triage
Which customers have the worst Account Health Check scores, and who slipped since the last refresh?	sophos-central-cli fleet health --limit 10
Where are the security and commercial gaps, customer by customer?	sophos-central-cli fleet gaps --kind all
Which machines have tamper protection off across every customer?	sophos-central-cli fleet tamper
Isolate this machine without looking up which tenant it is in	sophos-central-cli contain "WS-FIN-07" --scan --yes
What did each customer bill last month, and does it match what is deployed?	sophos-central-cli fleet revenue --month 2026-09
Feed every customer's SIEM events and alerts into my log pipeline	sophos-central-cli fleet siem-collect --stream both
Refresh every customer tenant into the local store	sophos-central-cli fleet refresh
Which credential am I using, and is it a partner, organization or tenant?	sophos-central-cli whoami

Full command reference: guide.md. For the AI-agent operating contract (--agent, --dry-run, when to confirm before mutating), see AGENTS.md.

What makes this different

Most Sophos Central integrations and MCP servers proxy each question into a live API call. That's fine for one record. It dies at scale, when you're asking "which of my 60 customers have machines with tamper protection off, and whose health score dropped since last week".

This skill syncs Sophos Central into a local SQLite mirror with full-text search. Aggregate questions become one local SQL join: instant, offline, and the AI sees the answer, not the raw data. Compound commands like fleet triage, fleet gaps and fleet revenue join across tenants, endpoints, alerts, Account Health Check scores, licences and partner billing - work a stateless API wrapper can't do.

Live wrappers such as rijul170/sophos-central-mcp take a tenant ID on each call and answer one tenant at a time. This connector also resolves the caller with whoami and routes every tenant call to that tenant's regional host for you, but fleet refresh first pulls every customer tenant into the local store with a per-tenant ok/forbidden/error ledger, so a missing customer shows up instead of silently dropping out of the answer. It complements the Sophos Central Partner Dashboard rather than replacing it: the dashboard stays the place for policy authoring, investigations and Sophos's own consolidated alert view.

The pain this closes

Sophos Central is tenant-by-tenant underneath. Sophos's own partner getting-started guide spells out the loop every MSP script has to run: whoami for the partner ID, page through the tenant list, then call each tenant's own regional API host with an X-Tenant-ID header. Community tooling usually stops at one tenant - the SophosCentral-Powershell module says "There is currently no support for multiple tenants" - so every MSP rewrites the same loop, and the questions behind the Partner Dashboard's alert roll-up still mean drilling into one customer at a time.

• sophos-central-cli fleet triage - one ranked morning work queue across every customer.
• sophos-central-cli fleet tamper - every machine with tamper protection off, previewed first, switched back on only with --yes.
• sophos-central-cli contain "WS-FIN-07" --scan - find a machine by hostname mid-incident without knowing its tenant and preview isolating it; add --yes to isolate.
• sophos-central-cli fleet revenue - monthly Sophos billing per customer next to what is actually deployed.

See pain-point.md for the longer narrative.

Frequently asked questions

Does this work with ChatGPT?

Not with this build. ChatGPT (Plus, Pro, Team, Business, Enterprise, and Education plans; Free does not yet expose Developer Mode) connects only to remote MCP servers over HTTPS, which means running the Sophos Central MCP server in HTTP mode. In this build HTTP mode checks a token but has no request size or time limits yet, so do not run it or put it behind a tunnel until the connector is reprinted; the security note in mcp-install.md explains why. Until then, use Claude Desktop, Claude Code or Codex.

Does this work with Codex, Cursor, Windsurf, Cline, Copilot, or Gemini?

Yes - all of them speak MCP. Cross-tool install commands are in the matrix above and the deep-dive in docs/which-agent.md.

Do I need to know how to code?

No. The recommended install is to paste one sentence into Claude Code or Codex - your agent reads SKILL.md and does the install. The fallback is a one-line installer per OS (bash or PowerShell). Neither path requires writing code. You'll enter your Sophos Central credentials once.

Is my Sophos Central data safe?

Your data stays on your machine. The CLI and MCP server are local binaries. The SQLite mirror sits in a directory under your user account. The AI agent only sees what the CLI returns - typically a query result, not raw bulk data. Credentials are read from your environment or your agent's config; never bundled into this repo or transmitted anywhere by MSP Skills.

What Sophos Central API credential do I need?

An API credential (a service principal) created in Sophos Central: Partner Dashboard > Settings & Policies > API Credentials for a partner, or Global Settings > API Credentials Management for an organization or a single customer. Set SOPHOS_CENTRAL_CLIENT_ID and SOPHOS_CENTRAL_CLIENT_SECRET. The role you give it is the real permission boundary: Service Principal Read-Only covers every report here but cannot run Live Discover queries; isolation, scans and tamper protection need a role that allows them.

Does it work with a partner, organization and single-tenant credential?

Yes. whoami tells the CLI which kind of credential it holds. A partner or organization credential walks every customer tenant in the fleet commands and picks one customer with --tenant (or SOPHOS_CENTRAL_TENANT_ID) for tenant-scoped commands such as endpoints or alerts. A single-tenant credential works on its own tenant.

Will this hit Sophos Central API rate limits?

Calls go through a shared throttle of 1.5 requests per second by default (--rate-limit), under Sophos's own limits. fleet refresh --plan prints the tenant count and estimated calls and time before you run it. After a refresh, fleet triage, fleet health and fleet gaps read the local store and make no API calls.

Why do some customers show as forbidden?

Sophos returns 403 to a partner for trial tenants and for term customers without Partner Assistance. fleet refresh records those in its per-tenant ledger and carries on, and fleet triage lists them as blind tenants so you can see what was not read. Trial tenants are skipped unless you pass --include-trial. A full refresh (no --tenant) also removes what earlier fleet refreshes stored for tenants it no longer covers (offboarded tenants, and trial tenants when --include-trial is not set), so fleet reports always match your last full refresh.

Can the AI isolate a machine or turn tamper protection back on by itself?

Only if you let it. contain and fleet tamper print a plan and change nothing without --yes, and the MCP server never passes --yes, so those two only preview over MCP. But the MCP server's generic sophos-central_execute tool can call any Sophos Central endpoint except the six that return a password or token, including isolation, scans and policy changes, gated only by your AI app's approval prompt. If an agent should only read, set SOPHOS_CENTRAL_MCP_READ_ONLY=1 in the MCP server's environment: it refuses every non-GET call and hides the sync and export tools. Keep the approval prompt on, and give the API credential a read-only role.

Does it cover Sophos XDR or MDR Cases and Detections?

Not in this version. It covers the REST APIs: alerts, endpoints, policies, exclusions, Live Discover, XDR Data Lake queries, SIEM events and alerts, Account Health Check, licences and partner billing. Cases and Detections are not included.

Does it replace the Sophos Central Partner Dashboard?

No. The dashboard stays best for policy authoring, investigations and day-to-day console work. This skill answers the across-every-customer questions and runs scripted actions from your AI agent.

What does it cost?

Free. Apache-2.0 licensed. You pay only for whichever AI agent you use (Claude, ChatGPT, Codex, etc.), and that's billed by your AI provider, not by us.

Safety model

Tier	Examples	Recommended agent policy
Read	fleet triage, fleet health, fleet gaps, fleet revenue, fleet refresh, whoami, tenants list, alerts list, endpoints list, search; contain and fleet tamper without --yes only preview	Allow
Endpoint response	contain --yes, endpoints set-isolation, endpoints bulk-isolate, endpoints scan, endpoints collect-forensic-log, endpoints collect-memory-dump, live-discover run-query, fleet tamper --yes	Human-in-the-loop, explicit confirmation - these reach real customer machines
Protection and policy config	exclusions add-scanning, allowed-items post-by-id, blocked-items post-by-id, policies update-policy, policies update-settings-of-policy, web-control update-web-decryption-excluded-sites, endpoint-groups add-multiple-endpoints-to-group	Preview with --dry-run, then a human-approved write - an exclusion or allowed item can open a hole
Write (routine)	alerts act-on, endpoint-tags, users create-tenant, health-check post-snooze, xdr start-query-run	Preview with --dry-run, then a reviewed write
Credential and access	endpoints get-tamper-protection, endpoints set-tamper-protection, migrations get-job (return a tamper password or migration token; kept out of the MCP server), auth login, partner-admins create, admins create-role-assignment	Human-in-the-loop only
Destructive and commercial	endpoints delete, endpoints bulk-delete, policies delete-policy, users delete-tenant-by-id, migrations create-sending-job, tenants create, tenants post-account-amendment	Human-in-the-loop only, explicit confirmation

Over MCP, the generic sophos-central_execute tool can reach every write in the table except the six that return a password or token; set SOPHOS_CENTRAL_MCP_READ_ONLY=1 for an agent that should only read. The strongest control is the scope you grant the Sophos Central credentials - the CLI can only do what the credentials are permitted to do. Full details, including how to lock it down, are in governance.md.

Status

Beta. Validated against the Sophos Central API surface and being validated with MSPs running it live against their own production tenant in our weekly Build Sessions. RSVP at compoundingteams.com/build-sessions.

---

Standards. Conforms to the open Agent Skills spec (Anthropic, Dec 2025; 40+ agents). MCP-compatible - works with any MCP-capable agent including Hermes. OpenClaw-ready (frontmatter pre-wired, awaiting OpenClaw launch).

Maintained by Servosity. Apache-2.0 licensed. Built with CLI Printing Press. Last updated: 2026-10-14.