jiradatacenter-mcp

jiradatacenter-mcp

MCP server for Jira Data Center / Server (self-hosted Jira, REST API v2)— exposes core work-item project-management operations (platform + Agile)as MCP tools.

MSPbots Connector gives your AI clients one MCP URL to reach your entire MSP stack (RMM, PSA, documentation, security, finance), and gives you control over who can call which tool and a record of everything that was called.

Hosted service: connect.mspbots.ai · Free to use.

Why Connector

Quick start

1. Sign up at connect.mspbots.ai.
2. Connect your vendors. Pick from 150+ connectors and authorize each one on the vendor's own page, or bring your own API keys. No extra vendor accounts needed.
3. Set access. Create roles and decide which tools each role may call. Nothing is allowed until you allow it.
4. Add your MCP URL to your AI client. Copy the URL from your Connector dashboard into Claude, ChatGPT, Cursor, VS Code, Gemini, Codex, or any MCP-capable client.
5. Ask your AI to do real work, and watch every call appear in the audit log.

Features

Unified gateway

• One MCP URL for all your AI clients
• 150+ connectors across the MSP stack
• Custom MCP servers supported, free
• Authorization happens on each vendor's own page; the AI never sees passwords or tokens
• Bring your own credentials: use the vendor API keys you already have
• Multiple accounts per app; connect once, reuse across AI clients
• Tool search: finds the right action for the task
• One sign-in; vendor credentials are injected behind the scenes and OAuth tokens refresh automatically

Visibility

• Audit log for every tool call (who, which vendor, when)
• Usage dashboard: call volume, top vendors, error rate, per-user activity
• Export audit logs to CSV, or stream to Loki, Graylog, or LogScale (SIEM)
• Optional recording of tool arguments and result summaries (off by default, enabled per organization)

Control

• Per-role, per-tool allowlist, deny by default
• owner / admin / member roles
• SCIM provisioning with Entra ID, Okta, JumpCloud, and Google Workspace

Security

• Credentials encrypted one record at a time (AES-256-GCM, per-record key)
• AI clients authenticate with OAuth 2.1 + PKCE
• Vendor responses are scanned for prompt-injection patterns and returned as structured data. This is one layer of defense, not a guarantee against every attack.
• Static egress IP for customer firewall and API allowlisting

Pricing

Free.

Links: connect.mspbots.ai · mspbots.ai

Overview

• Stateless HTTP service. No credentials are ever persisted — each requestsupplies its own site URL and Basic Auth credentials via headers, usedonly for the lifetime of that single request.
• Supports concurrent requests; per-request credential isolation is donevia Python contextvars, not a global/shared client instance.
• Entry points: POST /mcp (MCP protocol) and GET /health (health check).
• Default port: 8080 (configurable via MCP_HTTP_PORT).

Scope

15 tools, trimmed down from a prior 29-tool build (2026-08-07). The15 kept tools focus on core issue/search/project/user/agile-sprintoperations — full issue CRUD + comments + sub-tasks, JQL search, projectlookups, user lookups, and read-only agile board/sprint operations.Dropped in this trim: the niche metadata-lookup categories(issuetype, priority, jql autocomplete, customfields) and theless-common bulk/sync-oriented tools (version project-versionmanagement, worklog sync-since-timestamp endpoints), plus 2 of the 5agile tools (jira_agile_get_board, jira_agile_get_sprint — single-entity lookups that overlapped with the kept list-oriented board/sprinttools).

Everything else — customfields, issuetype, jql, priority,version, worklog (6 whole modules), plus jira_agile_get_board andjira_agile_get_sprint from agile.py — was removed from this build.

The original 214-tool full-WADL build was trimmed to 29 tools on2026-08-04. MSPbots' own stored integration config for this vendor is apre-generic-framework (legacy) integration — its 17 configured endpointentries return an entirely empty apiConfig (real endpoint logic lives inbackend Java code, not the generic config API), so there's no per-endpointmethod/path to reverse-engineer directly. The 17 endpoint names were usedas the category guide instead (Issue Search, Worklog, subtasks, User,request/Issue Comment, Project (Version), Board Sprint, Issuetype,Customfield, Priority, AutoCompleteData — "Queues"/"Queues Issue" map toJira Service Desk and "Tempo Worklog" to the third-party Tempo plugin,neither of which this REST-v2-platform-+-Agile-only server implements, sothose three are not covered), then for each matched category kept the oneor two operations that fit that name plus minimal same-category CRUD,capped at 30 total:

Everything else from the original 214-tool build — attachment, comment(the standalone properties-only category; real comment CRUD lives inissue), component, customfieldoption, dashboard, field, filter,group, groups, groupuserpicker, issuelink(Type), mypermissions,mypreferences, myself, permissions, projectcategory,projectvalidate, resolution, role, securitylevel, serverinfo,status, statuscategory (24 categories, ~184 tools) — was removed asunused by MSPbots and out of scope for this server's purpose.

Original sourcing (for context on where the removed categories came from,and where to look if one needs to be re-added):

• Platform API: parsed from Atlassian's official machine-readable WADLspec for Jira 9.4.0(https://docs.atlassian.com/software/jira/docs/api/REST/9.4.0/jira-rest-plugin.wadl).
• Agile API: manually curated from Atlassian's HTML docs athttps://docs.atlassian.com/jira-software/REST/9.4.0/ — no WADL ispublished for this API.

Authentication

Jira Data Center/Server uses HTTP Basic Authentication(Authorization: Basic base64(username:password)) — confirmed againstthe credentials MSPbots stores for this integration (its token field isliterally base64(username:password), not a separate secret). Thisserver builds that header itself from the raw username/password on everyrequest; the site's base URL is also per-tenant (each MSPbots tenant maypoint at a different Jira instance), so it is supplied as a header too,not a fixed environment variable.

HEADER authentication parameters

Missing any header returns 401:

{
  "error": "Missing credentials",
  "message": "This server requires the X-Jira-Site, X-Jira-Username and X-Jira-Password headers",
  "required_headers": ["X-Jira-Site", "X-Jira-Username", "X-Jira-Password"],
  "optional_headers": []
}

Environment Variables

MCP Endpoint

• POST /mcp — MCP protocol (streamable HTTP transport)
• GET /health — health check, returns {"status": "ok"} (pure local probe; does not depend on Jira availability)

Tool List

Tool names are jira_<category>_<operationId>, derived from the WADL<method id="..."> / manually-curated Agile operation_id. bodyparameters are accepted as a generic dict — see the Jira Server RESTAPI reference (linked below) for each resource's exact JSON schema.

Result Limits

Every list-style tool sends an explicit, clamped maxResults — none ofthem ever request an unbounded result set:

Responses are also capped at ~20,000 characters and truncated (with atruncated/original_count marker) rather than ever returned unbounded.

Test Example

# Health check
curl -s http://localhost:8080/health

# Call a tool via the MCP protocol (streamable HTTP) — requires an
# initialize handshake first per the MCP spec; abbreviated example below
# shows the tool-call request body only:
curl -s -X POST http://localhost:8080/mcp \
  -H "X-Jira-Site: jira.example.com" \
  -H "X-Jira-Username: api_user" \
  -H "X-Jira-Password: Sample_Passw0rd!" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -H "mcp-session-id: <session-id-from-initialize>" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "jira_myself_get_user",
      "arguments": {}
    }
  }'

Self-test limitation (2026-07-29): the provided test account(jira.mspbots.ai, user mspbots_api) could not be live-verified fromthis build environment — direct requests to jira.mspbots.ai (boththrough this server and via plain curl) consistently return a generic404 page not found, and DNS resolves the host to a sandbox-localaddress. This is consistent with MSPbots' internal servers beingRDP-only / network-isolated from this environment, not a bug in theserver itself. What was verified end-to-end from this environment:the MCP protocol handshake, tools/list returning all 214 registeredtools, the 401 credential-gating behavior, and that a live call reachesthe constructed URL and receives an HTTP response (proving therequest/auth-header plumbing works) — only the vendor-side reachabilityof the real instance could not be confirmed. A real functionalself-test against jira.mspbots.ai should be run from a network thatcan actually reach it (e.g. via RDP to an MSPbots-internal host).

API Reference

• Platform API WADL: https://docs.atlassian.com/software/jira/docs/api/REST/9.4.0/jira-rest-plugin.wadl
• Platform API docs: https://docs.atlassian.com/software/jira/docs/api/REST/9.4.0/
• Agile API docs: https://docs.atlassian.com/jira-software/REST/9.4.0/

Known Gaps

• Trimmed from 29 to 15 tools on 2026-08-07, focused on coreissue/search/project/user/agile-sprint operations. Dropped: the nichemetadata-lookup categories (issuetype, priority, jqlautocomplete, customfields) and the less-common bulk/sync tools(version project-version management, worklogsync-since-timestamp endpoints), plus 2 of 5 agile tools(jira_agile_get_board, jira_agile_get_sprint). See the Scopesection above for the exact category→tool mapping.
• Trimmed from 214 to 29 tools on 2026-08-04 (dropped the one write op, jira_agile_create_sprint, that wasn't part of MSPbots' read-only "Board Sprint" usage, to land under 30). The original buildexpanded beyond MSPbots' 17 named endpoints to full core work-itemcategories per an earlier scope decision. A later scope decision cutthis back down to what those 17 endpoint names actually map to, plusminimal same-category CRUD — see the Scope section above for the exactcategory→tool mapping and the full list of removed categories (24categories, ~184 tools). If a future need requires one of the removedcategories, Atlassian's WADL spec (linked above) still documents itsexact operations and they can be re-added the same way the kept toolswere generated.
• "Queues"/"Queues Issue" (Jira Service Desk) and "Tempo Worklog"(third-party plugin) are not covered at all — neither is part of theplatform-v2-REST-API-plus-Agile-API surface this server implements, so2 of MSPbots' 17 configured endpoint names have no corresponding toolhere even after the trim.
• body parameters are untyped (dict) rather than fully modeled —the Jira REST API reference documents the exact schema per resource,but reproducing all of them as typed Python parameters was out of scopefor a mechanically-generated server.
• Live functional self-test could not be completed from this buildenvironment due to jira.mspbots.ai being unreachable here (seeTest Example above for exactly what was and wasn't verified).
• Several kept operations mutate or delete real data(jira_issue_delete_issue, jira_issue_create_issue,jira_issue_edit_issue, jira_issue_add_comment, etc.) — treat anyDELETE/destructive POST/PUT tool as irreversible against aproduction Jira instance and confirm with a human before invoking.