| name | sc4sap:team |
| description | N coordinated SAP agents on a shared task list using Claude Code native teams |
| level | 3 |
| model | sonnet |
SC4SAP Team
Follows OMC team pattern adapted for SAP. Coordinates multiple specialized SAP agents working in parallel on a shared task list using Claude Code native team mode.
sc4sap:team spins up N coordinated agents that divide and conquer a SAP development task. Each agent is assigned a role (sap-developer, sap-code-reviewer, sap-transport-manager, etc.) and works on its slice of the task list simultaneously, with results merged at the end.
<Response_Prefix>
Every response triggered by this skill MUST begin with [Model: <main-model> · Dispatched: <sub-summary>] per ../../common/model-routing-rule.md § Response Prefix Convention.
</Response_Prefix>
<Phase_Banner>
Multi-phase skill. Before each Agent(...) dispatch (including every parallel spawn), emit ▶ phase=<id> (<label>) · agent=<name> · model=<Opus 4.7|Sonnet 4.6|Haiku 4.5> per ../../common/model-routing-rule.md § Phase Banner Convention.
</Phase_Banner>
<Use_When>
- Task has multiple independent SAP objects to create or modify (e.g., class + interface + unit test)
- User wants parallel development across multiple ABAP programs or packages
- Large transport release requiring parallel validation across multiple objects
- User says "team", "parallel", "all at once", or "split the work"
</Use_When>
<Do_Not_Use_When>
- Task is sequential (one object depends on another) -- use
/sc4sap:create-program (for a full program with ordered includes) or /sc4sap:create-object repeatedly
- Single object creation or modification -- delegate directly to executor
- User wants to review options first -- use
/sc4sap:deep-interview to crystallize the spec before dispatching agents
</Do_Not_Use_When>
<SAP_Agent_Roles>
Available roles to assign to team members:
sap-developer: Writes and activates ABAP code via MCP (CreateClass, UpdateClass, RunUnitTest, etc.)
sap-code-reviewer: Reviews ABAP code quality, naming conventions, and SAP best practices
sap-transport-manager: Handles CTS transport creation, assignment, and release via ListTransports, GetTransport, CreateTransport
sap-architect: Designs ABAP object structure, package hierarchy, and dependency graph
sap-tester: Writes and runs ABAP unit tests via CreateUnitTest, RunUnitTest, GetUnitTestResult
</SAP_Agent_Roles>
<Execution_Policy>
- Assign agents to non-overlapping task slices to avoid transport conflicts
- Each agent must confirm transport assignment before writing objects
- Merge results: collect activation status from each agent before team completion
- If any agent reports activation failure, halt and escalate to orchestrator
- Use
run_in_background: true for long-running unit test suites
</Execution_Policy>
1. Parse task into independent work slices (by ABAP object or package)
2. Assign one agent per slice with explicit role and task scope
3. Each agent: create/modify objects -> activate -> run unit tests -> report status
4. Orchestrator: collect all agent results, verify zero inactive objects remain
5. If all pass: report completion summary with object list and transport number
6. If any fail: surface errors and retry failed slices
Task: {{ARGUMENTS}}