Loading ToolBench
A collection of SAP development skills and templates for Claude Code, including validators and templates for ABAP, CAP, RAP, Fiori, HANA, Mobile, and S/4HANA extensibility
This server exposes 4 shell script wrappers as MCP tools with minimal metadata. Tool definitions lack proper JSON schemas, parameter type information, and descriptions are generic. No input validation, error handling guidance, or output schema documentation. Tools are inferred from shell scripts rather than explicitly registered with structured definitions. The server appears to be a skill collection installer rather than a functional MCP server.
Clean Core Compliance Checker - Scans ABAP/CDS code for Clean Core violations
ABAP Cloud Compliance Validator - Checks ABAP source files for non-released API usage and classic patterns
CAP Project Validator - Validates CAP project structure, CDS models, and configuration
RAP Business Object Validator - Checks RAP BO structure for completeness and best practices
No explicit tool registration visible. Tools are inferred from shell script filenames and descriptions. No JSON Schema input definitions found in source code.
Parameter descriptions are minimal (10-15 chars). 'File or directory path to validate (default: current directory)' lacks guidance on what validation checks occur, when to use this tool vs others, or what success looks like.
No output schema documented. Shell scripts return colored text with error counts, but LLMs cannot parse unstructured output. No structured response format (JSON, typed fields) defined.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 26 | 2026-07-28+ | v2 |
No error handling guidance. Shell scripts exit with code 0 or 1, but LLMs receive no actionable recovery steps. 'All checks passed' vs '5 errors, 2 warnings' lacks context on what to fix.
Tool names use hyphens (validate-abap-cloud) instead of underscores (validate_abap_cloud). Inconsistent with MCP naming conventions and harder for LLMs to parse.