Loading ToolBench
Real-time Multi-chain Wallet Intelligence Layer - MCP server for evaluating risks and verdicts for tokens and wallets on Solana and Robinhood Chain
Two tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. Both tools have descriptions (good), but descriptions are minimal (check_asset: 79 chars, check_wallet: 73 chars, below the 194-char baseline). Parameters lack detailed constraints and format guidance. No output schema documented. No error handling guidance. Tool names follow verb_noun pattern (check_*), which is acceptable but generic, 'evaluate_token_risk' and 'lookup_wallet_reputation' would be more specific. No tool annotations (readOnlyHint, etc.) despite both being read-only operations.
Evaluate risks and verdict for a token CA on Solana or Robinhood Chain.
Retrieve unified DNA reputation details for a wallet address.
Output schema not documented. LLMs cannot plan downstream calls or extract risk verdicts without knowing what fields are returned (e.g., risk_level, verdict, confidence_score, evidence).
Parameter descriptions lack format constraints and valid value guidance. 'chainId' description says '("solana" or "4663")' but should be an enum constraint in the schema, not prose. 'address' lacks format specification (Solana base58 vs EVM hex).
No error handling guidance. If an invalid address or unsupported chain is passed, the tool will fail silently or return a generic error. LLMs need recovery hints: 'Invalid address format. For Solana, use base58 (44 chars); for Robinhood, use 0x-prefixed hex (42 chars).'
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | 2026-07-28+ | v2 |
Tool descriptions are below baseline (45 chars vs 194-char average). 'Evaluate risks and verdict for a token CA' is vague, does it return a risk score, a boolean verdict, or a detailed report? 'Retrieve unified DNA reputation details' doesn't explain what 'DNA' means or what fields are included.
No tool annotations despite both tools being read-only. Adding readOnlyHint: true signals to agents that these calls are safe to retry and have no side effects, improving agent planning.