An unattended options trading agent that manages live positions, evaluates exit triggers, executes trades via Robinhood MCP, and performs post-exit tracking and self-reflection for day trades.
This server has critical definition gaps. All 4 tools lack visible input schemas in the source code, only descriptions are present in shell comments. Tool names use a namespace prefix (mcp__robinhood-trading__) but lack clear verb-noun structure (e.g., 'place_equity_order' is vague about order type, quantity, price). Descriptions are present but generic (50-80 chars), lacking WHEN to use, prerequisites, or return structure. No parameter documentation visible. No error handling guidance. No output schemas documented. The tools appear to be inferred from shell script comments rather than explicitly registered with full MCP schemas.
Cancel an equity order on Robinhood. Blocked in paper mode via --disallowedTools.
Cancel an options order on Robinhood. Blocked in paper mode via --disallowedTools.
Place an equity order on Robinhood. Blocked in paper mode and during autonomous runs via --disallowedTools.
Place an options order on Robinhood. Blocked in paper mode via --disallowedTools.
No input schemas visible in source code. All 4 tools lack documented parameter types, constraints, and descriptions.
Tool descriptions are generic and under 100 chars. Missing WHEN to use, prerequisites, return structure, and side effects. E.g., 'Place an equity order on Robinhood' does not explain required parameters (symbol, quantity, price, order_type) or what the response contains.
Tool names lack clear verb-noun structure. 'place_equity_order' is ambiguous, does it accept market/limit/stop orders? Does it require quantity, price, and symbol? Naming should be more specific (e.g., 'place_market_equity_order', 'place_limit_equity_order') or descriptions must clarify.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 31 | 2026-07-28+ | v2 |
No error handling guidance. Tools are destructive (place/cancel orders) but descriptions do not explain failure modes, retryability, or recovery steps. LLM has no guidance on what to do if an order fails.
Tools are inferred from shell script comments, not explicitly registered with MCP tool definitions. No visible MCP server code showing tool registration, schema binding, or handler implementation.