Loading ToolBench
The Brain Layer for AI Trading Agents — one MCP call returns a composite trade verdict (direction, confidence, regime) across major crypto perp venues. On-chain Merkle-verified track record.
Server has 7 tools with consistent naming (verb_noun pattern: get_*, scan_*) and complete input schemas with enums. However, descriptions are generic and repetitive, multiple tools share identical or near-identical descriptions ('Composite trade verdict...'), violating the pattern:tool-description requirement for distinct, actionable docstrings. Parameter descriptions are minimal (1-2 sentences) and lack context on when to use each tool or what distinguishes them. Output schemas are not documented in the source. No error handling guidance visible. Tool composition is reasonable (each does one thing), but descriptions fail to guide LLM selection between similar tools.
Equity trade verdict (direction, confidence, regime) for US equities and ETFs
Equity market regime classification (default symbol SPY)
Market regime classification on slower candles (1h, 4h, 1d)
Composite trade verdict (direction, confidence, regime) across major crypto perp venues
Composite trade verdict (direction, confidence, regime) across major crypto perp venues
Cross-venue funding arbitrage scanner
Market scanning tool for trade opportunities across venues
Duplicate/generic descriptions across tools. get_trade_signal and get_trade_call share identical descriptions ('Composite trade verdict...'), making it impossible for LLMs to distinguish when to use each. Violates pattern:tool-description requirement for distinct, context-specific docstrings.
Descriptions lack actionable context. No explanation of when to call each tool, what distinguishes crypto tools from equity tools, or how timeframe/exchange choices affect results. LLMs cannot infer selection logic from generic 1-2 sentence descriptions.
Output schemas not documented. No visible documentation of what fields each tool returns, their types, or structure. LLMs cannot plan downstream calls or extract required data without knowing response shape.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 61 | 2026-07-28+ | v2 |
scan_funding_arb has minimal schema. Only 'symbol' parameter with no enum constraints or format guidance. Unclear what symbols are valid (BTC, ETH, etc.) or if arbitrary strings are accepted.
No error handling guidance visible. No documentation of failure modes, retryability, or recovery steps. LLMs cannot self-correct or know whether to retry on failure.