Loading ToolBench
MCP server for HoodScope — Robinhood Chain deployer reputation, insider detection, and token verdicts. Paid tools auto-pay via x402 (USDC on Base).
HoodScope defines 6 tools with clear, domain-specific descriptions (avg 180 chars, within baseline 194). All tools have input schemas with type declarations and parameter descriptions. However, naming lacks action verbs (hoodscope_* prefix is domain-specific but not verb-driven), and output schemas are completely undocumented, LLMs cannot predict response structure for chaining. Error handling exists but is generic ('HoodScope error'). No tool annotations (readOnlyHint, destructiveHint). All tools are READ_ONLY by design, which is correct, but this should be declared via annotations.
Full reputation rap sheet for a deployer wallet on Robinhood Chain: how many tokens it launched, how many reached real activity vs. died, best-token volume, whether it does wallet-to-wallet insider distribution, and a 0-100 score. Costs $0.01 USDC on Base.
Poll the newest Robinhood Chain token launches, each auto-scored by its deployer's reputation (trusted / mixed / avoid / serial_spammer / new_deployer). Use min_score=70 to see ONLY launches from proven deployers (the high-signal alpha — ~1% of launches), since=<cursor from a prior call> to get only new launches, limit=N (max 100). Designed to be polled. Costs $0.01 USDC on Base.
The highest-reputation deployers on Robinhood Chain (score ≥ 70), ranked. Costs $0.01 USDC on Base.
Free coverage stats for HoodScope: how many deployers are scored, tokens indexed, real tokens, insider wallets flagged, and serial-spam factories detected on Robinhood Chain.
On-chain activity for a Robinhood Chain token: swap count, WETH volume, unique traders, whether it crossed the real-activity threshold, and its deployer's score. Costs $0.01 USDC on Base.
No output schemas documented. LLMs cannot predict response structure (fields, types, nesting) for downstream chaining or data extraction. All 6 tools return JSON but structure is invisible to the agent.
Tool names lack action verbs. 'hoodscope_verdict' does not start with get_, search_, or list_. LLMs infer intent from verb-noun naming; domain prefixes alone are weaker signals for tool selection.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). All tools are read-only queries, but this is not declared in the schema. Agents cannot distinguish safe-to-retry tools from stateful ones without annotations.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 61 | 2026-07-28+ | v2 |
Composite BUY-RISK verdict for a Robinhood Chain (chain 4663) memecoin: combines the deployer's full-history reputation, insider-distribution flags, and on-chain activity into a single signal (trusted / mixed / avoid / serial_spammer / new_deployer) with plain-English reasons. Works even on a token that launched seconds ago, because it scores WHO deployed it. Costs $0.02 USDC on Base.
Generic error messages. 'HoodScope error {status}' and 'HoodScope call failed' do not guide recovery. No distinction between retryable (network timeout) vs user-fixable (invalid address) vs fatal errors.
Parameter descriptions lack format/constraint details. 'token' and 'address' are described as '0x…' but no validation rules (length, checksum, chain-specific format) are stated. LLMs may pass invalid addresses.