Loading ToolBench
MCP Server for Business Central Cursor rules and editor configuration
This server has severe definition quality issues. Only 1 tool is defined (bc_help), with minimal documentation and no real implementation. The tool has a basic schema but lacks critical depth. The server appears to be a proof-of-concept or template rather than production-grade. Parameter descriptions are absent, output schemas are undocumented, and error handling is non-existent. The tool definition exists but is functionally incomplete.
Get help with Business Central development
Tool name 'bc_help' lacks action verb and is vague. Does not follow verb_noun naming convention (e.g., get_help, search_help, generate_help). LLMs cannot infer what happens when called.
Tool description 'Get help with Business Central development' is generic and under 30 characters effective meaning. Missing: WHAT help is provided, WHEN to call this vs other discovery tools, WHAT data format is returned, any prerequisites.
Parameter 'topic' has a description but no type constraint, enum, or format specification. The description 'Topic to get help with' is trivial and does not guide LLMs. No examples of valid topics, no constraints on length or format.
Output schema is completely undocumented. Tool response is hardcoded as { result: 'This is a sample result' } with no type information. LLMs cannot plan downstream operations without knowing the response structure.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 0 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 26 | - | v1 |
No error handling or recovery guidance. If the tool fails or receives an invalid topic, there is no actionable error message or classification (retryable vs fatal). Error path returns generic console.error.
Tool implementation is a stub returning hardcoded sample data. Does not actually provide help on any Business Central topic. Code shows: const result = { result: 'This is a sample result' }, placeholder, not functional.
No pagination or result limiting mechanism. If this tool were to return help topics or documentation, there is no limit parameter, offset/page support, or guidance on maximum result size. Large unstructured responses would blow context window.
Tool composition is incomplete. The server defines cursor rules (naming conventions, error handling, patterns) but the single tool does not help agents apply or validate those rules. No tools to validate naming, suggest patterns, or generate code examples based on Business Central domain.