Orchestrator/.kiro/steering/mcp-servers.md

inclusion

inclusion
auto

MCP Server Configuration

Available MCP Servers (require VPN)

These MCP servers are hosted on DB's OpenShift cluster (dbcs-prag) and available when connected to the corporate network:

Jira + Confluence (Atlassian MCP — NEUER kombinierter Server)

• URL: https://taros-playground-mcp-atlassian.apps.dbcs-prag.comp.db.de/mcp
• Capabilities: 72+ Tools — Jira issues, transitions, comments + Confluence pages (read/write)
• Headers required:
  • Authorization: Bearer <JIRA_TOKEN>
  • X-Jira-Host: arija-jira.jaas.service.deutschebahn.com (ohne https://)
  • X-Jira-Projects: O2CO2C,O2CAAPN,* (komma-separiert, Wildcards möglich)
  • X-Jira-Mode: read (oder write)
• Replaces: mcp-jira-v2 (deprecated) and mcp-confluence (separate)
• Use for: Team velocity, sprint data, bug tracking, Confluence pages

Jira MCP v2 (DEPRECATED)

• URL: https://taros-playground-mcp-jira-v2.apps.dbcs-prag.comp.db.de/mcp
• Status: DEPRECATED — use mcp-atlassian instead

Confluence MCP (standalone, still works)

• URL: https://taros-playground-mcp-confluence.apps.dbcs-prag.comp.db.de/mcp
• Still functional but mcp-atlassian is the preferred replacement

DB Context (dbctx)

• URL: https://taros-playground-dbctx.apps.dbcs-prag.comp.db.de/mcp
• Capabilities: DB-specific context and tooling
• Use for: General DB knowledge, terminology, organizational context

Teamlandkarte MCP (local, stdio)

• Repo: https://git.tech.rz.db.de/ThomasHandke/teamlandkarte-mcp
• Type: Local stdio server (Python, requires uv + Data Lake credentials)
• Capabilities: Match tasks to employees/teams by competences, browse free capacities, team profiles
• Use for: Finding teams (INKA, Data Science), capacity analysis, competence mapping
• Requires: DATA_LAKE_USERNAME, DATA_LAKE_PASSWORD, AZURE_OPENAI_EMBEDDING_API_KEY
• NOT yet configured (needs credentials)

BEAM MCP (local, stdio) — EDUCATIONAL USE ONLY

• Repo: https://git.tech.rz.db.de/beta/ai/mcp/beam-mcp
• Type: Local stdio server (Node.js, requires Edge browser for login)
• Capabilities: Enterprise Architecture (LeanIX/BEAM) — applications, domains, components, diagrams
• Use for: System landscape, application dependencies, architecture diagrams
• ⚠️ ACTIVE USE NOT PERMITTED due to data protection concerns. Educational reference only.

Other Noteworthy Servers (not yet configured)

Server	Use Case	Priority
DataLake MCP	SQL on Trino (quantitative data)	Medium — when we need metrics
Foundations MCP	ADRs, Standards, Capabilities	Medium — for architecture comparison
gitlab-mcp	GitLab integration via glab	Low — we use direct API calls
sonarqube-mcp	Code quality metrics	Low — for team quality comparison
proclib-mcp	Process library access	Medium — for process analysis
mcp-agilehive	AgileHive SAFe tool	High — PI-Planning data, team capacity

Fallback (without VPN)

When MCP servers are unreachable, use direct REST API calls:

• Confluence: Bearer token auth against https://arija-confluence.jaas.service.deutschebahn.com/rest/api
• GitLab: PRIVATE-TOKEN header against https://git.tech.rz.db.de/api/v4
• Jira: Not available without VPN

Local MCP Server

• OrgMyLife: Running locally, accessible via Kiro's MCP integration
• Provides: tasks, calls, digest, projects board