fin-advisor
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PYTHONPATH | Yes | Path to the AlphaDesk repository (e.g. C:\AlphaDesk) so that the finagent module can be found when running python -m finagent.mcp_server. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
美股投研工具集（NVDA/MSFT/AAPL/TSLA/AMZN）。行情为日线收盘价。财报问答请先调用 search_filings 并用 [chunk_id] 引用。draft_order 只生成模拟订单，必须由用户在本地运行 python -m finagent.orders confirm <order_id> 人工确认，模型不能代替用户确认，也不存在任何真实交易通道。
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_quoteB | 查询最新日线收盘价、涨跌幅、52 周高低点（非实时）。ticker: NVDA/MSFT/AAPL/TSLA/AMZN/QQQ |
| get_price_historyB | 查询最近 N 个交易日收盘价和区间涨跌幅。 |
| get_company_newsB | 查询公司近期新闻标题（Yahoo Finance RSS 缓存）。 |
| search_filingsA | 在最新 10-K/10-Q 原文中做 BM25 检索，返回带 chunk_id 的段落。回答时用 [chunk_id] 引用。query 建议用英文。 |
| get_key_financialsB | 查询 SEC XBRL 结构化财务数据（revenue, net_income, eps_diluted, operating_cash_flow, capex 等）。 |
| portfolio_riskA | 计算组合年化波动率、最大回撤、HHI 集中度、相对 QQQ 的 beta。holdings 例: [{"ticker":"NVDA","weight":0.6}] |
| draft_orderB | 生成模拟订单（PENDING_CONFIRMATION）。不执行真实交易。需用户在本地 CLI 人工确认。 |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 7 tools
Most tools are cleanly separated by resource and action: snapshot quote vs. time-series history vs. news vs. unstructured filing search vs. structured XBRL financials. The only mild overlap is get_quote vs. get_price_history, both returning close prices, though the 52-week/snapshot framing versus range-history framing makes the boundary reasonably clear.
Six of seven tools follow a snake_case verb_noun pattern (get_quote, get_price_history, get_company_news, search_filings, get_key_financials, draft_order). portfolio_risk is a noun-only name that breaks the verb convention, but the overall scheme remains highly predictable.
Seven tools is well-scoped for a research-plus-advisory assistant: each one maps to a distinct research or action capability. Nothing feels redundant or padded, and nothing essential to the described workflow is missing purely for count reasons.
The surface covers the full research-to-decision lifecycle: market data, history, news, filings, structured financials, portfolio analytics, and a simulated order draft with a safe human-confirmation boundary. Minor gaps exist around order lifecycle (no listing, cancel, or status of draft orders) and multi-ticker screening, but agents can work around these.