Loading ToolBench
Alpha MCP exposes 16 tools with consistent naming (verb_noun pattern: get_*, predict_*, trigger_*) and reasonable descriptions (avg ~120 chars). All tools have input schemas with type definitions and parameter descriptions. However, output schemas are undocumented, tools return JSON strings without declaring field structure. Error handling is minimal: predict_kronos catches exceptions but most tools lack recovery guidance. No tool annotations (readOnlyHint/destructiveHint). Symbol validation via _check_symbol() is good security practice. Missing: pagination support, per-item error reporting, and structured output documentation.
获取 Hermes Agent 运行状态：是否在运行、LLM 可用性、最近任务。
获取策略漏斗快照：candidate/focus/buy 三个池子的股票列表和统计。
获取当日热门概念板块 Top 10。
获取当日热门个股 Top 30。
获取指定个股的日 K 线数据。
获取 K 线缓存统计：总股票数、覆盖天数分布。
获取 K 线缓存同步状态：最后同步时间、覆盖率、进度。
Output schemas undocumented. All tools return JSON strings without declaring field structure, types, or required fields. LLMs cannot plan downstream calls or extract data reliably.
No tool annotations. Tools lack readOnlyHint/destructiveHint/idempotentHint. Clients cannot determine which tools are safe to retry or have side effects.
Minimal error recovery guidance. predict_kronos catches exceptions but returns raw error text. Most tools lack actionable error messages or recovery hints.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 63 | 2026-07-28+ | v2 |
获取公告个股详情：公告内容 + K 线。
获取公告选股漏斗：三个池子的公告股票列表和统计。
获取公告选股的关键词规则列表：标签和权重。
获取个股详情：基本面指标 + K 线数据。
获取个股盘中实时行情（当天 OHLCV、涨跌幅、成交额）。
获取当前策略配置：所有可调参数及其当前值。
列出最近的 Agent 任务执行记录。
调用 Kronos 时序模型预测个股未来 K 线走势。返回历史+预测合并 K 线、预测起始索引、预测明细。 这是真实的 AI 模型推理结果，任何关于股价走势的预测必须通过本工具获取，禁止自行编造预测数据。
触发公告筛选：从当日公告中筛选利好个股。
No pagination support. Tools returning lists (get_hot_stocks, list_agent_tasks) lack limit/offset and total count. Large result sets risk context window exhaustion.
Parameter constraints not enforced. trade_date, days, kline_days, lookback, horizon lack min/max bounds. LLMs can pass absurd values (e.g. days=999999) causing API failures.