Fortinet MCP

Fortinet + 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 Fortinet API. Not affiliated with, endorsed by, or sponsored by Fortinet, Inc.

_{Full skill page - install, outcomes, safety model.}

Read every FortiGate's configuration and live state, then audit the whole client fleet for firmware, SSL-VPN exposure, admin accounts and license expiry in one command. 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 read your data 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 Fortinet skill
Claude Desktop	Run installer, then Settings > Extensions to register fortinet-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 fortinet-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 Fortinet MCP (.mcpb) - then open Claude Desktop > Settings > Extensions and select the file. One click, no JSON, no shell. (Browse every Fortinet 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 fortinet@msp-skills

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

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

Install the Fortinet 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/fortinet/install.sh). If it uses Windows PowerShell, run iwr -useb https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/fortinet/install.ps1 | iex. Then authenticate per the README and run fortinet-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/fortinet/install.ps1 | iex

macOS / Linux:

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

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

Verify:

fortinet-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/fortinet/install.sh)

Windows (PowerShell):

iwr -useb https://raw.githubusercontent.com/Servosity/msp-skills/main/skills/fortinet/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 fortinet@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/fortinet --force

Inside a Hermes chat session:

/skills install servosity/msp-skills/skills/fortinet --force

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

Install for OpenClaw

Tell your OpenClaw agent (copy this):

Install the fortinet skill from https://github.com/servosity/msp-skills/tree/main/skills/fortinet. The skill defines how its required CLI (fortinet-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

Create a REST API admin on the FortiGate (System > Administrators > Create New > REST API Admin, read-only profile, your IP as a trusted host), then:

FORTINET_BASE_URL=<value> FORTINET_API_TOKEN=<value> fortinet-cli doctor

doctor checks config and paths, then makes one authenticated read (GET /monitor/system/status) so it reports whether the token actually works, not just whether one is set. For a fleet of FortiGates, list them in a devices.toml file and run fortinet-cli fleet devices to see which entries have a usable URL and token file.

What this skill does

Question your MSP keeps asking	Command
One posture verdict across every client FortiGate, only findings new since the last run	fortinet-cli fleet audit --new
Which FortiGates are below the fixed firmware for their branch?	fortinet-cli audit firmware --min-version 7.2.10,7.4.5
Where is SSL-VPN on a WAN interface and open to any source?	fortinet-cli audit sslvpn
Which admins have no trusted hosts or no two-factor?	fortinet-cli audit admins
Which FortiGuard and FortiCare licenses expire in the next 60 days?	fortinet-cli audit licenses --within 60d
Which firewalls have HA out of sync, tunnels down, or certificates expiring?	fortinet-cli fleet health --cert-within 14d
Which policies are any-any accept, unused, or not logging?	fortinet-cli policies hygiene --unused-days 90
What changed in each client's configuration this week?	fortinet-cli fleet changes --since 7d
Which FortiManager-managed devices are out of sync?	fortinet-cli fmg devices
Which FortiGate Cloud devices have the tunnel down or a contract ending?	fortinet-cli cloud devices

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

alpadalar/fortigate-mcp-server is a general-purpose FortiGate MCP server: a device registry, read tools, and create, update and delete tools for policies, VIPs and routes that stay off unless you enable writes. It answers about the device you ask about.

This one is read-only by construction and built for the question an MSP asks the morning an SSL-VPN advisory lands: which of my 40 clients are exposed right now? A devices.toml file lists every client FortiGate, and fleet audit reads all of them in parallel, then grades each one PASS, WARN, FAIL or UNKNOWN for firmware floors per branch, SSL-VPN exposure, admin accounts and license expiry. A box it could not read is UNKNOWN, never PASS. Every run is recorded locally, so fleet audit --new shows only findings that are new since the last one, and fleet backup plus fleet changes show which configuration objects moved between two backups. fmg and cloud add the FortiManager and FortiGate Cloud inventory alongside, also read-only. Stored secrets stay out of the answers: every FortiOS-encrypted value comes back as ENC (redacted).

Fortinet's FortiAI-Assist runs inside FortiManager and FortiAnalyzer, and since FortiOS 8.0 on FortiGates with more than 2 GB of RAM and FortiCare Premium; it answers inside the console it runs in. This needs only a read-only REST API token per FortiGate, so it also covers the small boxes FortiAI-Assist does not, and it answers across every client's firewall at once.

The pain this closes

FortiOS SSL-VPN is the part of a client's network that gets attacked first. When Fortinet disclosed CVE-2024-21762 in February 2024 it said the flaw was "potentially being exploited in the wild" (Rapid7), and the advice for MSPs was to move every box to the fixed release of its own branch or turn SSL-VPN off (SmarterMSP). Doing that the same morning means knowing which client FortiGates run SSL-VPN, which listen on WAN open to anyone, and which are below the fixed build. FortiOS answers one box at a time; without a FortiManager, that is a login per client.

• fortinet-cli audit sslvpn - FAIL for SSL-VPN on a WAN interface open to any source, per device and VDOM.
• fortinet-cli audit firmware --min-version 7.2.10,7.4.5 - every box below its branch floor, and how far behind.
• fortinet-cli audit licenses --within 60d - one renewal queue instead of a spreadsheet.
• fortinet-cli fleet audit --new --fail-on fail - all four checks, new findings only, exit 6 for a scheduled job.

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 Fortinet 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 Fortinet credentials once.

Is my Fortinet 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.

Can it change my FortiGates?

No. Every FortiOS call is a read of configuration tables or monitor state (the settings and monitor commands read; they do not set), the FortiManager commands send only the JSON-RPC get method, and the FortiGate Cloud commands only list devices. Stored secrets stay out of the answers too: every FortiOS-encrypted value (RADIUS secrets, IPsec keys, password hashes) comes back as ENC (redacted), and a request for plain-text passwords is refused. Give each token a REST API admin with a read-only admin profile and the FortiGate itself enforces the rest.

How do I check 40 client FortiGates, not just one?

List them in devices.toml, one [[device]] entry each with its url, a token_file (a file holding that box's token; on macOS and Linux the CLI refuses it unless it is chmod 600) and an optional client name. fortinet-cli fleet devices shows which entries have a usable URL and token file, and every audit command then runs across the whole list. A devices file you name (with --devices or FORTINET_DEVICES) that is missing or lists no [[device]] entries is an error, never a silent fall-back to one box. Without one, the audits check the single FORTINET_BASE_URL device.

My FortiGate still uses its factory certificate. Will it connect?

Not by default: certificate checks stay on. The secure fix is a trusted certificate on the FortiGate (FortiOS can get one from Let's Encrypt). Otherwise opt in per device with insecure = true in devices.toml, pass --insecure, or add "FORTINET_SKIP_TLS_VERIFY": "true" to the MCP server's env block in a JSON config (it also turns verification off for FortiManager; FORTINET_FMG_INSECURE=true covers FortiManager alone). The one-click .mcpb has no field for it on purpose, because one ticked box would turn off verification for every device. A devices.toml entry with insecure = true covers the fleet, audit and policies hygiene tools there; the single-device tools and the fmg tools still need a trusted certificate (or the JSON config above).

Do I need FortiManager or FortiGate Cloud?

No. Both are optional extras: the fmg commands read FortiManager ADOMs and managed devices, and the cloud commands read your FortiGate Cloud inventory. Everything else talks to each FortiGate directly.

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 audit, fleet health, the four audit checks, policies hygiene, fmg and cloud reads, every list / get, monitor and settings command, sync, search, export (encrypted secrets redacted; sync and fleet audit write local state, export --output overwrites the file you name)	Allow
Sensitive read	fleet backup (full configurations incl. encrypted secrets and private keys, unredacted on purpose), administrators, api-users, local-users, monitor fortitokens	Human-in-the-loop; keep output private
Credential	auth set-token, auth logout	Human-in-the-loop only
Data egress	--deliver webhook:<url>, feedback --send	Human-in-the-loop

No command changes a FortiGate, FortiManager or FortiGate Cloud.

The strongest control is the admin profile you give each REST API token - a read-only profile means the FortiGate refuses anything else. Full details, including how to lock it down, are in governance.md.

Status

Beta, awaiting live verification. The command surface is generated from a community-authored OpenAPI description of FortiOS 7.4 endpoints, derived from Fortinet's public Ansible collection (not an official Fortinet specification), and the fleet audits are covered by unit tests against simulated FortiGate, FortiManager and FortiGate Cloud responses, but no MSP has yet run it against production FortiGates and reported back - so nothing here is proven against real devices. If you run it against your own, the it-works form is how that becomes a badge. MSPs put connectors like this through their paces 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: 2026-10-13.