abap-mcp

MCP server for SAP ABAP — offline static analysis, ABAP Cloud readiness, and RAP scaffolding. No SAP system. No credentials. Works on ABAP source wherever your AI agent works: a git checkout, an abapGit export, a code review, CI.

Built on abaplint (the open-source ABAP parser/linter) and the Model Context Protocol. TypeScript, 100% local — the server makes zero network calls and touches no filesystem: sources go in as text, findings come back as structured JSON.

Why this exists

Every other ABAP MCP server is either a bridge to a live SAP system (ADT/RFC — needs credentials, a system, and trust) or a documentation search. But AI coding agents spend most of their time where the files are — editing abapGit repos, reviewing diffs, generating code — long before anything reaches a system. This server gives agents the missing feedback loop at that layer:

• "Does this ABAP parse? Is it clean?" → lint_abap

• "How far is this classic report from ABAP Cloud?" → check_cloud_readiness

• "Start me a correct RAP business object." → scaffold_rap_bo

• "What's in this 4,000-line class?" → get_abap_outline

Related MCP server: ABAP-ADT-API MCP-Server

Quickstart

# Claude Code
claude mcp add abap-mcp -- npx -y abap-mcp

# or any MCP client (.mcp.json / mcp.json):
{
  "mcpServers": {
    "abap-mcp": { "command": "npx", "args": ["-y", "abap-mcp"] }
  }
}

From a clone instead:

npm install && npm run build
claude mcp add abap-mcp -- node /path/to/abap-mcp/dist/cli.js

Then ask your agent things like "lint this class against ABAP Cloud", "is zold_report cloud-ready?", or "scaffold a RAP BO for entity Booking on table zbooking, draft enabled".

Tools

Honesty box — what this is not

• Not ATC. Readiness here is language-level: statements ABAP Cloud removed. Whether your code calls released APIs only requires a system's released-API list (ATC check SAP_CP_READINESS) — out of scope for an offline tool, and the readiness report says so on every call.

• Scaffold validation is tiered. Generated classes and CDS views are round-tripped through abaplint at Cloud level before they're returned (the generator and the linter share one parser). Behavior/service definitions are outside abaplint's checked surface — they are golden-tested canonical templates, and ADT activation is the final arbiter. Each generated file is labeled validated: "abaplint" | "template".

• Text-in only, by design. No filesystem walking, no network — the entire attack surface is a parser over strings you explicitly pass. For linting whole directories, use the abaplint CLI in CI, or the mcp-kit wrap-abaplint recipe this server grew out of.

Develop

npm install
npm run check     # typecheck + 80 tests + build — the CI gate
node dist/cli.js  # stdio MCP server
npx @modelcontextprotocol/inspector --cli node dist/cli.js --method tools/list

Tool descriptions are CI-graded (a rubric test enforces verb-first names, when-to-use, non-goals, described params, worked examples — the mcp-kit discipline; the full mcp-kit lint scores all seven tools 100/100).

Design

The decision log — why offline, why abaplint, why a dual-parse readiness diff, why the scaffolder validates its own output, what was deliberately left out — lives in docs/DESIGN.md.

Credits

• abaplint by Lars Hvam — the parser and rule engine underneath every tool here (MIT).

• mcp-kit — the production-MCP patterns this server follows (typed tool specs, transport discipline, description lint).

MIT © Akshay Palimkar. Not affiliated with or endorsed by SAP SE. "SAP", "ABAP" and "RAP" are trademarks of SAP SE; this is an independent open-source tool for developers working with them.