Loading ToolBench
MCPB Desktop Extension for Indian rent-control drafting. Orchestrates a multi-agent pipeline for drafting legal pleadings in Indian rent control cases.
Three tools with clear names and detailed descriptions, but critical schema gaps. list_case_types and get_case_type_format have empty input schemas ({}), violating the requirement that every parameter be typed and described. get_agent_instructions has one optional parameter with a description but no type constraint. Descriptions are verbose (200+ chars) and informative but lack the concise LLM-optimized format (target 50-200 chars). No output schemas documented for any tool. Tool annotations present (readOnlyHint, destructiveHint, idempotentHint) but incomplete, missing openWorldHint on some tools. No error handling guidance or recovery paths defined.
Get the instructions for the drafting pipeline. Default mode (no agent_name, or agent_name=""): returns the FULL orchestration script enumerating every step and tool call from case-folder creation through final-draft.docx. THIS IS THE MANDATORY FIRST CALL when the user asks you to draft anything. Single-agent mode (agent_name ∈ {reader, format, drafter, verifier, refiner, overseer}): returns that agent's persona instructions only.
Get the drafting format SKILL.md for a given case type. Returns the full SKILL.md content for the case type, including statutory anchors, format placeholders, and the drafting checklist. Also appends format-from-user.md if present in the same skill folder.
List all case types this connector can draft. Returns a dict with the list of case types, a short description of each, AND an acronym disambiguation map (Indian advocacy acronyms / Form references / statutory section references -> case_type). CRITICAL: the user's acronym IS the source of truth for case_type classification. Do not infer case_type from input-document content.
list_case_types has empty input schema ({}). Tool is read-only but provides no structured output schema documentation.
get_case_type_format accepts case_type parameter with description but no type constraint (should be enum of valid case types). LLM cannot validate input without explicit enum.
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields without knowing return structure. Descriptions mention 'returns dict' or 'returns content' but do not specify fields.
Descriptions are 200+ characters, exceeding LLM-optimized range (50-200 chars). Verbose descriptions waste tokens and bury key intent. E.g., list_case_types description is 280+ chars.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 50 | 2026-07-28+ | v2 |
No error handling guidance. If case_type is invalid or skill file missing, no recovery path documented. LLM receives raw error with no actionable next step.