Generate a complete RAP stack from a short spec: CDS root view entity → behavior definition (managed) → behavior implementation class → service definition → service binding. DEFAULTS TO dryRun:true — returns the planned object names and generated source for review WITHOUT creating anything. Set d...
AI agents use adt_rap_scaffold to create or update resources in Claude For Abap — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Claude For Abap environment.
This tool creates or modifies data (ABAP source objects) in a live SAP system. Although it defaults to dryRun:true for safety, it is fundamentally a creation tool capable of reversibly adding multiple objects to the system. The blast radius is significant (creates entire RAP stacks) but not permanent or financial, making it Write rather than Destructive.
From the tool's definition Tool description states: 'Set dryRun:false to actually create them in dependency order' and 'Built on the same create primitives as adt_create_object.' It generates and creates multiple ABAP objects (CDS root view entity, behavior definition, behavior…
Risk signalsAdmin/system-level operation
Documented attack patterns abuse exactly the kind of access adt_rap_scaffold gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Claude For Abap, and nothing reaches the server without passing your rules. This is the rule we recommend for adt_rap_scaffold:
{
"version": "1",
"default": "deny",
"tools": {
"adt_rap_scaffold": {
"limits": [
{
"counter": "adt_rap_scaffold_rate",
"window": "minute",
"max": 30,
"scope": "grant"
}
]
}
}
}adt_rap_scaffold stays usable, but capped — an agent stuck in a loop can't make hundreds of changes a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Generate a complete RAP stack from a short spec: CDS root view entity → behavior definition (managed) → behavior implementation class → service definition → service binding. DEFAULTS TO dryRun:true — returns the planned object names and generated source for review WITHOUT creating anything. Set dryRun:false to actually create them in dependency order (subject to read-only mode; requires a transport for non-local packages). The service binding is always plan-only (its publish step is not automated). Built on the same create primitives as adt_create_object. It is categorised as a Write tool in the Claude For Abap MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Claude For Abap MCP server in PolicyLayer and add a rule for adt_rap_scaffold: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Claude For Abap. Nothing to install.
adt_rap_scaffold is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the adt_rap_scaffold rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for adt_rap_scaffold. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
adt_rap_scaffold is provided by the Claude For Abap MCP server (yzonur/claude-for-abap). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Claude For Abap, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
45 Claude For Abap tools catalogued and risk-classified — across an index of 43,000+ MCP servers.