Autotask MCP

Autotask + AI - for Claude, the ChatGPT desktop app, GitHub Copilot, Gemini CLI, and any agent that speaks MCP

Unofficial. Community-built Claude Code Skill and MCP server for the Autotask API. Not affiliated with, endorsed by, or sponsored by Datto, Inc.

[Autotask PSA demo - animated preview]

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

Every Autotask entity at the command line, plus a local SQLite mirror that answers ticket-aging, workload, unbilled-time, and account-360 questions no other Autotask tool can. Works with the AI you already use - ChatGPT (desktop app), Claude Desktop, Codex, Claude Code, Claude Cowork, and GitHub Copilot - plus Gemini CLI. 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 has no password, listens on every network interface, and predates the browser-rebinding guard, so a web page you open can reach it. 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):

The ChatGPT desktop app runs autotask-mcp locally and shares MCP configuration with Codex (OpenAI). ChatGPT on the web needs a remote (HTTP) server, and this connector's HTTP mode is paused until it is reprinted (see the security note in mcp-install.md), so msp-skills does not support it yet. If your team works there, a hosted gateway is the better pick; check that your AI app is on the gateway's setup list.

Also for the Microsoft and Google stacks

Big install base, but an honest heads-up: Gemini CLI runs the local binary you just installed, while Microsoft 365 Copilot and the Gemini app need a remote server, which msp-skills does not support yet.

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

Claude Desktop bundle: the .mcpb launches on macOS (Intel and Apple Silicon), Windows x64 and Linux x64. Every tool shells out to the companion autotask-cli. Releases cut from 2026-09-18 on ship that CLI inside the bundle; older bundles contain only the MCP server, so if yours lacks the companion, run the Path A or Path B installer below first (or set AUTOTASK_CLI_PATH to an existing binary). Details: #331.

Download Autotask MCP (.mcpb) - then open Claude Desktop > Settings > Extensions and select the file. One click, no JSON, no shell. (Browse every Autotask 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 autotask@msp-skills

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

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

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

macOS / Linux:

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

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

Verify:

autotask-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/autotask/install.sh)

Windows (PowerShell):

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

Add to Claude Desktop, the ChatGPT desktop app, GitHub Copilot, Gemini CLI, 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 ChatGPT desktop app. 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/autotask --force

Inside a Hermes chat session:

/skills install servosity/msp-skills/skills/autotask --force

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

Install for OpenClaw

Tell your OpenClaw agent (copy this):

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

AUTOTASK_API_INTEGRATION_CODE=<value> AUTOTASK_PSA_SECRET=<value> AUTOTASK_PSA_USER_NAME=<value> autotask-cli doctor

doctor confirms the credentials work before you run anything that touches data. It exits 0 even when the credential is rejected, so scripts should add --fail-on error.

What this skill does

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 Autotask 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 "how much approved time is unbilled across every client this month" or "which retainers run out before the next renewal" - each one a multi-call dance against a per-tenant zone URL, every categorical filter gated by integer picklist IDs that vary per instance, and all of it spending your hourly API request budget.

This skill syncs Autotask 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 reconcile, retainer, and company-360 join across tickets, time entries, contracts, invoices, and resources - work a stateless API wrapper can't do.

The pain this closes

Ask any MSP owner where Autotask hurts and you get two answers: time that never gets billed, and reporting that means an export. Techs close tickets without entering time, or approved hours never make it onto an invoice, and nobody catches it until the billing run - because Autotask has no single screen joining approved time to the invoices it belongs on. And when you want the cross-object picture, Autotask's native LiveReports is a copy-and-edit report designer that outputs to Excel, PDF, RTF, or CSV (per Kaseya's own Autotask PSA reporting help) - so anything joining tickets to time to contracts gets pushed into Power BI or another BI tool. An entire third-party connector market exists for exactly that.

This skill answers those questions where the data already lives - one offline query, no report build, no spreadsheet:

See pain-point.md for the longer narrative.

Frequently asked questions

Does this work with ChatGPT?

Yes, in the ChatGPT desktop app. It runs MCP servers locally (STDIO), the way autotask-mcp runs, and shares MCP configuration with Codex (OpenAI): add autotask-mcp under Settings > MCP servers as a STDIO server, or add it once to Codex's ~/.codex/config.toml. ChatGPT on the web is not supported by msp-skills yet: it needs a remote (HTTP) server, and this connector's HTTP mode is paused until it is reprinted; the security note in mcp-install.md explains why. If your team works there, a hosted gateway is the better pick; check that your AI app is on the gateway's setup list.

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 Autotask credentials once.

Is my Autotask 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 credentials do I need?

An API-only user created in your own Autotask instance, which gives you a UserName and Secret, plus the tracking Integration Code from your integration vendor record. Set AUTOTASK_PSA_USER_NAME, AUTOTASK_PSA_SECRET, and AUTOTASK_API_INTEGRATION_CODE in your environment. The API user's security level is the real permission boundary - scope it to exactly what you want the AI to reach.

Do I have to figure out my Autotask zone URL?

No. Autotask hosts each tenant in a numbered zone (webservicesN.autotask.net). Run autotask-cli zone once and the CLI discovers and caches your tenant's base URL - the same zone-information request Autotask's own docs tell integrations to make first.

Will this hit my Autotask API rate limits?

Autotask meters API calls per database in a rolling 60-minute window and adds latency as you approach the threshold. The local mirror exists so reads stop hitting the API: after the first sync, the cross-entity views (unbilled, reconcile, retainer, company-360, ticket-aging, triage, workload) run against local SQLite with zero API calls, and sync is incremental - it only fetches what changed since the last checkpoint.

Is this Datto Autotask PSA or just Autotask?

Same product. Autotask PSA is now branded Datto Autotask PSA under Kaseya; this skill targets the Autotask REST API, which both names refer to. It does not cover Datto RMM - that's a separate product with a separate API.

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

The strongest control is the scope you grant the Autotask 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.

Credential safety

• The three Autotask headers (ApiIntegrationCode, UserName, Secret) and every API request go only to the scheme, host and port of the configured base URL (AUTOTASK_BASE_URL, or the zone URL autotask-cli zone saved). A redirect to any other host is refused, so the credentials never reach it. zone asks only webservices.autotask.net and saves only an https://webservicesN.autotask.net/atservicesrest zone URL.
• Any output line or JSON document, file write, local store row or MCP result that would repeat a credential this program holds (AUTOTASK_API_INTEGRATION_CODE, AUTOTASK_PSA_SECRET, AUTOTASK_MCP_HTTP_TOKEN, the stored integration code, or one in config.toml) is withheld whole, in any encoding, and "[credkit: withheld from stdout: it carried a credential this program holds]" appears in its place (naming stderr, mcp, file, sqlite, child process or outbound request where that is the sink). A credential-named field in a response (a webhook secretKey, a client portal user's password, an integration vendor's secret) is replaced by [withheld], so the rest of the record stays readable. Responses are never cached on disk.
• The integration code and the Secret are the credentials this program holds. The API user name (AUTOTASK_PSA_USER_NAME, sent as UserName) is the account's login name and is not treated as a secret: an answer that repeats it or its email domain (Resources, Contacts) prints and syncs as it comes.
• Only credentials.toml (mode 0600) holds a credential; auth set-token reads the integration code from stdin, never from an argument.
• --dry-run sends nothing. Over MCP, a tool that changes something (in Autotask, or this CLI's saved zone) returns a preview and acts only with confirm: true, and a tool argument can never name a file to read or write.
• A --stdin body is held as a credential, whole and line by line: a preview of it (--dry-run) and an answer that repeats part of it are withheld, while a write still sends it. feedback takes its note as an argument, not from stdin.
• config.toml never keeps the userinfo or query of base_url; they stay in AUTOTASK_BASE_URL.
• While any credential is configured, agent-context (one large JSON document) prints only the withheld notice; run it with no credential configured (for example a fresh AUTOTASK_HOME and the credential variables unset) to read it.

Status

Beta. Validated against the Autotask 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: 2026-06-06.