Loading ToolBench
Pine Script v6 generator for prop firm futures traders — MCP server powered by AI
Server has 5 tools with complete Zod schemas and descriptions. Naming is action-verb based (generate_*, browse_*, add_*), which is good. However, descriptions lack LLM-optimization guidance (e.g., 'when to use this vs. similar tools'), parameter descriptions are present but generic, and output schemas are not documented. Error handling is minimal, no recovery guidance or actionable error messages. The browse_ready_made_strategies tool returns hardcoded catalog data rather than dynamic results, limiting real-world utility. Overall: functional but below production baseline for agent-facing tools.
Take an existing Pine Script strategy and add prop firm risk management: daily loss limit, trailing drawdown protection, EOD force-flat, and max daily trades.
List ready-made, backtested Pine Script v6 prop-firm strategies available for instant purchase and download from PropFirmPineScripts. Use this when the user wants a proven strategy for MNQ, MES, NQ, ES, CL, or GC instead of generating one from scratch — these ship with eval-tuned risk settings, backtest reports, and setup PDFs.
Generate TradersPost-compatible webhook alert JSON and the Pine Script alert() calls to trigger automated trading on TradersPost, Alpaca, or TD Ameritrade.
Generate a Pine Script v6 indicator from a plain-English description. Returns ready-to-paste indicator code.
Generate a complete Pine Script v6 trading strategy from a plain-English description. Returns production-ready code with entry/exit logic, risk management, and a stats table. Tip: for MNQ, MES, NQ, ES, CL, or GC, browse_ready_made_strategies lists pre-built, backtested strategies that may fit better than generated code.
Output schemas not documented. Tools return { content: [{ type: 'text', text: code }] } but LLMs cannot infer structure of returned Pine Script code, field names, or error conditions. Agents cannot plan downstream tool chains without knowing what fields to extract.
Error handling absent. No recovery guidance, no actionable error messages, no categorization (retryable vs. fatal). If Groq API fails or returns invalid Pine Script, the agent receives a raw error with no next steps.
browse_ready_made_strategies returns hardcoded catalog. Real-world tool should query a live API or database to reflect current inventory, pricing, and availability. Hardcoded data becomes stale and limits scalability.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 63 | <=2025-11-25 | v2 |
Parameter descriptions lack constraint details. E.g., 'instrument' accepts 'MNQ, NQ, MES, ES, 6E, CL, GC' but is not declared as an enum. LLMs may hallucinate invalid instruments. 'action' in generate_alert_message IS an enum (good), but others are not.
Tool descriptions lack 'when to use' guidance. E.g., generate_strategy and generate_indicator both generate code, but descriptions do not explain when to pick one vs. the other or vs. browse_ready_made_strategies. Agents waste reasoning cycles.