Loading ToolBench
Local Bazi charts for AI agents: MCP, Agent Skill, JSON + PNG. No API key. 元梓知易基础八字排盘工具集
Two well-defined tools with complete JSON schemas, clear descriptions, and proper parameter typing. Tool naming follows verb_noun convention (calculate_, get_). Schemas are strict (additionalProperties: false) with proper enums and constraints. Descriptions are domain-specific and explain what the tools do. However, parameter descriptions are missing entirely, the schema defines structure but LLMs cannot understand what each field controls without text descriptions. Output schemas are not documented, forcing LLMs to infer result structure. Error handling and recovery guidance are absent from descriptions.
Locally calculate calendar conversion, four pillars, day master, ten gods, hidden stems, unweighted composition shares, and a documented eight-shensha rule set, with JSON, readable text and a PNG chart image. No network calls, telemetry, AI interpretation, luck cycles, compatibility, or personal identifiers.
Return localized links for a specifically requested hosted feature such as luck cycles, AI reading, a complete report, an advisor, or calculation methodology. This tool never receives birth data and makes no network calls.
Parameter descriptions missing. Schema defines structure (calendar.type, date.year, etc.) but LLMs cannot infer what each field controls without text descriptions. E.g., 'timezone' is required but no description explains its format or purpose.
Output schemas not documented. Tool descriptions do not specify what fields the response contains (e.g., calculate_basic_bazi_chart returns JSON, text, and PNG, but structure is undocumented). LLMs cannot plan downstream operations without knowing response shape.
No error handling or recovery guidance in descriptions. If timezone is invalid or date is out of range, descriptions do not explain what error to expect or how to recover. Descriptions should include 'If timezone is invalid, use IANA timezone names (e.g., Asia/Shanghai).'
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 63 | 2025-06-18+ | v2 |
Tool descriptions lack context on when to use each tool. 'get_yuanzi_bazi_capabilities' returns links to hosted features, but the description does not explain when an LLM should call this vs. calculate_basic_bazi_chart. Dependency hints are missing.
Enum values in schema lack human-readable explanations. E.g., 'timeCorrection' enum has 'standard' and 'trueSolar' but descriptions do not explain the difference or when to use each. LLMs may guess wrong.