name: mulesoft description: "MuleSoft integration suite orchestrator. Routes MuleSoft work to the right specialist subskill — connector config, flow construction, DataWeave transforms, platform ops, or BDR project tracking. Use as entry point when a MuleSoft request spans multiple areas or when not sure which subskill fits. NOT for a single narrow task — use the specific subskill directly."
Skill: MuleSoft Suite Orchestrator
Purpose
Entry point for MuleSoft integration work. Reads the request, identifies which subskill(s) apply, loads them, and coordinates cross-skill work. Use when the request spans multiple areas (e.g. "build the Phase 1A flow end-to-end" needs connector + flow + dataweave + platform).
When to Use
- Request spans multiple MuleSoft areas
- Not sure which subskill fits
- Starting a new integration project from scratch
- Orchestrating a full phase (design → build → test → deploy)
When NOT to Use
- Single narrow task — go direct to the subskill
- BDR project status / blockers — use
/mulesoft-bdrdirectly
Subskill Map
| Skill | Responsibility |
|---|---|
/mulesoft-bdr |
BDR project tracker, playbook state, blockers, stakeholder comms |
/mulesoft-connector |
Connector config (SF, NS, HTTP, DB, JMS, FTP), auth patterns |
/mulesoft-flow |
Flow construction, triggers, error handling, idempotency |
/mulesoft-dataweave |
DataWeave 2.0 transforms, field mapping |
/mulesoft-platform |
Anypoint Platform ops, deploy, monitoring, Secrets Manager |
/mulesoft-agentfabric |
AI & agents layer — Agent Fabric, Omni Gateway, MCP tools, LLM Proxy, Agentforce bridge |
Routing Logic
| Request pattern | Subskill |
|---|---|
| "Where am I on BDR?" / "What's blocked?" | mulesoft-bdr |
| "Set up / configure a connector" / "Test connection" | mulesoft-connector |
| "Build / debug a flow" / "Add error handling" | mulesoft-flow |
| "Map these fields" / "Transform JSON to XML" | mulesoft-dataweave |
| "Deploy" / "Secrets Manager" / "CloudHub alerts" | mulesoft-platform |
| "AI agents" / "MCP server/tool" / "Omni Gateway" / "LLM proxy" / "Agentforce" / "agent network" | mulesoft-agentfabric |
Multi-Step Workflows
Workflow: New Integration (end-to-end)
mulesoft-bdr— confirm phase gate passedmulesoft-connector— add connector, store credentialsmulesoft-flow— build flow structure with trigger + error handlermulesoft-dataweave— write transforms for field mappingmulesoft-platform— deploy to Design, then Sandbox, then Productionmulesoft-bdr— update phase status, notify stakeholders
Workflow: Debug Production Issue
mulesoft-platform— check Runtime Manager logs, error rate, last runmulesoft-bdr— is there a related blocker (credential rotation, etc.)?mulesoft-connector— is the connection still valid?mulesoft-dataweave— if data error, check transform logicmulesoft-flow— if flow error, check error handler coverage
Workflow: Phase Gate Check
mulesoft-bdr— read gate criteria for current phase- Check each criterion against actual project state
- Report pass/fail with specific missing items
- Route unblocking tasks to relevant subskill
BDR Programme Context
Phase 1 scope (Account Suspension):
- Source: NetSuite Customer
entityStatus - Target: Salesforce Account
On_Stop__cflag - Pattern: 15-min polling, watermarked, idempotent upsert
- Timeline: Build 21-25 Apr → Test 28 Apr-2 May → Prod 5 May 2026
Future phases (per Enterprise strategy docs Phase 1-6):
- Phase 2: Broader enterprise integration architecture
- Phase 3: Use case catalogue
- Phase 4: Multi-system implementation
- Phase 5: Governance & observability
- Phase 6: AI & advanced automation →
/mulesoft-agentfabric(Agent Fabric / Agentforce). Two hard gates first: BDR entitlement for Agent Fabric (NOT confirmed in "Integration Advanced" tier) + a concrete agent use case.
Core Principles (apply across all subskills)
- Environment promotion is Design → Sandbox → Production. Never shortcut.
- Credentials in Secrets Manager, never plaintext. Especially production.
- Every flow is idempotent. Retries safe. No duplicate writes.
- Every external call has retry + error handler. No silent failures.
- Every polling flow uses watermarks. Object Store tracks last-run time.
- Every production deploy is tagged in Git. Rollback must be trivial.
- Every integration ships with Runtime Manager alerts. Unmonitored = broken.
Proactive Triggers
- Request mentions "production" without "JWT auth" mentioned → flag auth upgrade
- Request mentions "deploy" without mentioning UAT sign-off → flag promotion risk
- Request mentions "new connector" without mentioning Secrets Manager → flag credential handling
- Request spans 3+ subskills → suggest breaking into sequential sub-tasks via this orchestrator
Anti-Patterns (do NOT do these)
- Skipping straight to a subskill when the request actually needs orchestration across 3+ skills
- Loading all 5 subskills at once — only load what's needed for the current step
- Trying to answer MuleSoft questions without grounding in the playbook
- Inventing MuleSoft features — verify with official docs or the Expert Coaching session
Related Skills
- All
mulesoft-*subskills (routed to above) /sync-knowledge-base— after updating any skill, sync to GitHub + Notion/project-refresh— if BDR project context seems stale