One-stop data layer for AI Agents. Finance: A-share stock/index/ETF data, market overview, multi-dimensional screening, universal search across 11,780 securities, gold price. Info: web search, news, trending topics, daily bulletin. Life: weather, calendar (lunar/solar terms/holidays/trading days).
GroundAPI MCP server demonstrates solid definition quality with 11 well-named, action-oriented tools covering finance, info, and life domains. All tools have descriptions (avg 150 chars) and input schemas with typed parameters. However, output schemas are undocumented, LLMs cannot predict response structure for downstream chaining. Parameter descriptions are present but inconsistent in depth; some lack constraints (e.g., 'limit' params missing min/max bounds). Error handling is basic (ToolError wrapping) without recovery guidance. Tool names follow verb_noun convention well (finance_stock, info_search). No tool annotations (readOnlyHint, destructiveHint) despite all being read-only operations.
Real-time gold price in CNY/gram and USD/oz.
Market-wide A-share data — for 'what's happening in the market' questions.
Find stocks matching user criteria. ALL params optional — combine 1-3 you need.
Universal search across 11,780 A-share securities by code, name, pinyin, or ticker.
Query A-share securities (stocks/indices/ETFs). One stop for everything about a security.
Daily market bulletin: key events, announcements, and market summary.
Output schemas completely undocumented. All 11 tools return dict but LLMs cannot predict field names, types, or structure for response parsing and downstream chaining.
Numeric parameters (limit, days, pe_min, pe_max, etc.) lack explicit min/max bounds in descriptions. LLMs may pass invalid values (e.g., limit=1000 when API caps at 50).
No tool annotations despite all tools being read-only. Adding readOnlyHint=true would signal to agents that these calls are safe to retry and have no side effects.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 64 | <=2025-11-25 | v2 |
Financial news from major Chinese news outlets (新浪财经, 东方财富, 同花顺, etc.).
Web search across news, blogs, forums, and general web content.
Trending topics and hot searches from Weibo, Baidu, and other platforms.
Chinese calendar with lunar dates, solar terms, holidays, and trading days.
Weather forecast and current conditions for any Chinese city.
Error handling is minimal. ToolError wrapping provides no recovery guidance. When API returns 404 or validation error, LLM receives no hint about what to try next (e.g., 'Try search_finance() first').
Parameter descriptions for 'aspects' (finance_stock) and 'scope' (finance_market) use prose lists instead of enums. LLMs may hallucinate invalid values like 'aspects=earnings_forecast' instead of using documented options.