name: mulesoft-platform
description: "Anypoint Platform operations expert. Covers Runtime Manager (deploying, monitoring, alerts), Secrets Manager, Exchange (assets, templates), environment promotion (Design → Sandbox → Production), Access Management (roles, permissions), vCore allocation, Object Store, Runtime Fabric, and CI/CD via Maven. Use for any platform admin task, deployment, monitoring setup, or governance configuration."
Skill: Anypoint Platform Operations
Purpose
Operate Anypoint Platform like a platform admin — deploy apps, manage credentials, configure monitoring, promote across environments, and govern integrations at scale. Everything beyond mvn deploy and before "it's running in prod".
When to Use
- Deploying an app to Design / Sandbox / Production
- Configuring Secrets Manager for credentials
- Setting up Runtime Manager alerts
- Managing user access and environment permissions
- Downloading or publishing Exchange assets
- Setting up Maven deploy in CI/CD
- API Manager policies (Client ID Enforcement, rate limiting) — see
references/api-manager-policies.md
- CH1 vs CH2 deploy distinction (CH2 has NO Runtime Manager Promote) — see
references/deployment.md
When NOT to Use
- Writing connector configs — use
/mulesoft-connector
- Building flows — use
/mulesoft-flow
- DataWeave transforms — use
/mulesoft-dataweave
- BDR project tracking — use
/mulesoft-bdr
- Omni Gateway / Agent Fabric / MCP / LLM Proxy / Agentforce — use
/mulesoft-agentfabric. Omni Gateway is Flex Gateway extended for agentic traffic; it is NOT the classic Anypoint Platform admin covered here. Classic REST API Manager policies still live in this skill.
The Platform Stack
Anypoint Platform (eu1.anypoint.mulesoft.com)
├── Access Management ← Users, Teams, Business Groups, Environment permissions
├── Runtime Manager ← Deploy, monitor, alert, vCore allocation
├── Secrets Manager ← Credential storage per environment
├── Exchange ← Assets, templates, connectors, APIs
├── API Manager ← API gateway, policies, analytics
├── Design Center ← API design (RAML, OAS)
├── Monitoring ← Dashboards, logs, metrics
└── Object Store ← Shared key-value store for watermarks
Non-Negotiable Rules
- Production credentials live ONLY in Secrets Manager. Never in source code, property files, or deploy commands.
- Environment promotion is Design → Sandbox → Production. No shortcuts. Every app passes through all three.
- Every production app has Runtime Manager alerts configured — app stopped, error rate, flow unresponsive.
- Never deploy straight to production without UAT sign-off in Sandbox.
- Production deploys happen during low-traffic windows unless absolutely critical.
- Every deploy produces a tagged Git commit for rollback.
Modes
| Mode |
When |
Reference |
| Deploy app |
New deploy or update existing |
references/deployment.md |
| Configure Secrets |
Before first deploy to each env |
references/secrets-manager.md |
| Set up monitoring |
After production deploy |
references/monitoring.md |
| Manage access |
Grant/revoke roles, add users |
references/access-management.md |
| Environment promotion |
Design → Sandbox → Prod |
references/environments.md |
| Maven / CI/CD |
Automated deploys |
references/maven-cicd.md |
Core Concepts
| Term |
Meaning |
| Organization |
Top-level account (BDR Group) |
| Business Group |
Sub-org for multi-tenant setups |
| Environment |
Design / Sandbox / Production (BDR has 3) |
| Worker |
Unit of CloudHub runtime (0.1, 0.2, 1, 2 vCores) |
| vCore |
Virtual CPU allocation for a worker |
| Runtime |
Mule Runtime version (4.11.x) |
| App |
Deployed integration project |
| Asset |
Exchange item — connector, template, API, custom |
| Secret Group |
Collection of secrets in Secrets Manager |
BDR Environment Model
Confirmed from Access Management screenshot:
| Environment |
Purpose |
CloudHub Admin |
vCores Allocated |
| Design |
Dev, individual work |
Adam (granted) |
TBC |
| Sandbox |
UAT, pre-prod testing |
Adam (granted) |
TBC |
| Production |
Live workloads |
Adam (granted) |
TBC |
vCore allocation — check Runtime Manager → Usage. BDR has Premier plan, specific allocation to be confirmed.
Proactive Triggers
- App deployed to production without Runtime Manager alerts configured → flag immediately
- Credentials in plaintext
config-production.yaml → flag as security risk
- CloudHub deploy using
--skipTests=true → flag as quality gap
- App running on 0.1 vCore in production → flag as undersized
- Multiple apps sharing the same Secret Group → flag as blast radius risk
- Production deploy without corresponding git tag → flag for rollback difficulty
- User asks to BUILD a custom monitoring/observability Mule app → STOP and check their Anypoint subscription tier FIRST. Most enterprise tiers (Integration Advanced, Titanium, Gold) include Anypoint Monitoring + Functional Monitoring, which cover ~95% of monitoring use cases natively. Building custom monitoring code without confirming the subscription is wasted work. See
references/monitoring.md decision tree.
- Before recommending vCore scale-up or new app deploy → check Subscription Usage panel for current cap headroom. Overage charges apply on most plans. See
references/monitoring.md "Object Store + usage caps" section for the screens to check.
Templates
| Template |
Purpose |
templates/maven-settings.xml |
Maven settings with Anypoint repository auth |
templates/cloudhub-deploy.yaml |
CloudHub deploy config |
templates/alerts-config.md |
Recommended Runtime Manager alerts |
templates/github-actions-deploy.yml |
GitHub Actions workflow for CI/CD |
Anti-Patterns (do NOT do these)
- Storing production credentials in property files committed to git
- Deploying directly to production without passing through Sandbox UAT
- Sharing one Connected App across all environments (SF)
- Using "Share everything with everyone" in Access Management
- Skipping Test Connection before deploy
- Deploying without a rollback plan (git tag, previous worker version)
- Running production on 0.1 vCore "to save money" — it'll OOM under load
Related Skills
- Use
/mulesoft-connector for Secrets Manager secrets setup
- Use
/mulesoft-flow for the app being deployed
- Use
/mulesoft-dataweave for transforms inside the app
- Use
/mulesoft-bdr for BDR-specific deploy coordination