Jira Service Management MCP

Jira Service Management + 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 Jira Service Management API. Not affiliated with, endorsed by, or sponsored by Atlassian Pty Ltd.

[Jira Service Management demo - animated preview]

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

Every Jira Service Management Cloud request-API (servicedeskapi) and Assets endpoint in one binary, plus SLA breach ranking across queues, internal notes that stay internal, ticket timelines, and Assets (CMDB)-to-ticket joins. 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 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):

The ChatGPT desktop app runs jira-service-management-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

Download Jira Service Management MCP (.mcpb) - then open Claude Desktop > Settings > Extensions and select the file. One click, no JSON, no shell. (Browse every Jira Service Management 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 jira-service-management@msp-skills

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

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

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

macOS / Linux:

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

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

Verify:

jira-service-management-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/jira-service-management/install.sh)

Windows (PowerShell):

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

Inside a Hermes chat session:

/skills install servosity/msp-skills/skills/jira-service-management --force

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

Install for OpenClaw

Tell your OpenClaw agent (copy this):

Install the jira-service-management skill from https://github.com/servosity/msp-skills/tree/main/skills/jira-service-management. The skill defines how its required CLI (jira-service-management-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 an Atlassian API token at id.atlassian.com/manage-profile/security/api-tokens, signed in as an agent on your JSM site. Then set three variables: your site subdomain (acme for acme.atlassian.net), your Atlassian email, and the token.

macOS / Linux:

export JIRA_SERVICE_MANAGEMENT_SITE=acmeexport [email protected]export JIRA_SERVICE_MANAGEMENT_API_TOKEN=<api-token>jira-service-management-cli doctor

Windows (PowerShell):

$env:JIRA_SERVICE_MANAGEMENT_SITE = "acme"$env:JIRA_SERVICE_MANAGEMENT_EMAIL = "[email protected]"$env:JIRA_SERVICE_MANAGEMENT_API_TOKEN = "<api-token>"jira-service-management-cli doctor

Assets (CMDB) commands also need your site's cloudId and Assets workspaceId (Assets is part of JSM Premium and Enterprise). Run jira-service-management-cli cmdb discover --save once per client site to look both up and store them in the config file for that site; saved ids apply only to the site they were found for. As a backstop, every Assets call is refused unless its cloudId is the current site's (checked once per run against the site's public tenant_info) and its workspaceId is one the site lists, so a leftover JIRA_SERVICE_MANAGEMENT_CLOUD_ID or _WORKSPACE_ID from another client can never point Assets at that client's CMDB. Environment values win over saved ids, so unset them when you switch sites.

doctor checks config and paths, names any missing SITE, EMAIL or API_TOKEN, and verifies the token with one authenticated read (GET /servicedesk).

What this skill does

client and sla-report read the local copy: run jira-service-management-cli sync --resources organization,request,requesttype,servicedesk first. Everything else reads live.

Through the MCP server, cmdb reconcile reads an export only from inside the server's working directory, and refuses to run when that directory is a filesystem root or your home folder (so it cannot be used to read arbitrary files). If your MCP host starts servers from / or your home folder, run cmdb reconcile from the CLI in the folder that holds the export instead.

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 Jira Service Management integrations and MCP servers proxy each question into a live API call. That's fine for one record. It breaks down when you ask "which of our 300 open tickets breach SLA next": no single API call ranks tickets across queues, so the AI has to walk every queue, drop duplicates, and read each ticket's SLA metrics itself, one tool call at a time.

This skill does that work inside the binary. breaches walks every queue live, de-duplicates tickets and ranks them by SLA time remaining; request timeline merges comments (labeled public or internal), status history, approvals and SLA state into one view; cmdb tickets joins Assets (CMDB) objects to every ticket connected to them. For period questions it also keeps a local SQLite copy with full-text search: after a sync, sla-report, client and search answer from the local copy with no API calls.

Atlassian's official Rovo MCP Server is hosted by Atlassian and covers Jira and Confluence plus a growing set of JSM tools: request search, create, transition and approvals (developer docs), operations alerts and Assets search (support docs). The community mcp-atlassian server covers Jira and Confluence broadly, including service desk queues and internal comments. This connector complements them: it runs on your machine, ranks tickets by JSM's own SLA timers across queues, keeps a local copy for period SLA reports and per-client views, and diffs Assets against your RMM or PSA export.

The pain this closes

Internal notes reach customers by accident. Atlassian's own public tracker carries the request (JSDCLOUD-17210): agents "don't know if they are making an internal or external facing comment and may accidentally have selected the wrong one." An AI agent posting through a generic comment call has the same problem. Meanwhile no single API call ranks open tickets across queues by SLA time left, so "what breaches next" is a queue walk plus a per-ticket SLA read, and Assets (CMDB) only knows what someone imported, so drift against your RMM shows up during an outage.

• jira-service-management-cli request note SD-123 "..." - always posts public=false and errors if JSM stores the comment as public.
• jira-service-management-cli request timeline SD-123 - every comment labeled PUBLIC or INTERNAL, before you reply.
• jira-service-management-cli breaches --within 2h - every queue walked once, tickets de-duplicated and ranked by SLA time remaining.
• jira-service-management-cli roster 3 - client staff in no organization or in two, and organizations not attached to the desk.
• jira-service-management-cli cmdb reconcile --type Laptop --file rmm.csv --key Name - Assets diffed against an RMM or PSA export.

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 jira-service-management-mcp runs, and shares MCP configuration with Codex (OpenAI): add jira-service-management-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 Jira Service Management credentials once.

Is my Jira Service Management 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.

Does this work with Jira Service Management Data Center or Server?

No. It targets Jira Service Management Cloud (your-site.atlassian.net). Data Center and Server use different hosts and authentication, and this connector does not call them.

Do I need JSM Premium for Assets (CMDB)?

Only for the Assets commands. Assets is part of JSM Premium and Enterprise. Queues, SLAs, requests, comments, customers and organizations work on any JSM Cloud plan. For Assets, run jira-service-management-cli cmdb discover --save once per client site: it finds that site's cloudId and Assets workspaceId and stores them for that site only.

Will an internal note ever reach the customer?

Use request note for anything internal: it always posts with public=false, and if JSM reports the stored comment as public it exits with an error so the agent knows. request comment add is the customer-reply command: pass --public for a customer-visible reply. It sends no public field when you omit the flag, so never use it for internal notes. Read request timeline before replying: it labels every comment PUBLIC or INTERNAL.

Will this hit my Atlassian API rate limits?

Live commands make normal API calls and pace themselves to the rate-limit headers Atlassian returns (--rate-limit overrides it). breaches reads each queue and then each ticket's SLA, so scope it with --desk and --queue on large sites. sla-report, client and search read the local copy after a sync and make no API calls.

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

Writes send immediately: --dry-run is an opt-in preview, not a default. Set JIRA_SERVICE_MANAGEMENT_MCP_READ_ONLY=1 in the MCP server's environment and it drops request_note, sync, export, workflow_archive, learnings_forget, learnings_reject and cmdb_reconcile (which reads a local file); its generic jira-service-management_execute tool refuses any non-GET API call, and cmdb_discover refuses --save. The local learning tools (teach, teach_*, playbook_amend, learnings_confirm, recall) stay: they write only to the CLI's own local store (recall logs each lookup there), never to your site.

The strongest control is the Atlassian account behind the API token - the CLI can only do what that account is permitted to do, so create the token as an agent with the access you want the AI to have. Full details, including how to lock it down, are in governance.md.

Status

Beta. Validated against the Jira Service Management 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: 2026-10-16.