Loading ToolBench

Reading the current corpus…

chenswap-mcp MCP server · D · ToolBench

chenswap-mcp

MCP server for chenswap.chitacloud.dev — compare DEX quotes across 5 aggregators (KyberSwap, ParaSwap, Odos, LI.FI, Bebop) with LLM reasoning. No API key, no wallet connection, read-only quotes.

2026-07-28+STDIO2 tools

54/100

Definition Quality74

Protocol Readiness40

Supportability31

Spec posture

Inferred effective spec 2026-07-28+·current baseline 2026-07-28

Static source inference · medium confidence

Current-spec patterns detected

Summary

Two tools with complete input schemas and descriptions. get_quote has detailed parameter documentation (6 params, all typed and described); list_supported_chains is minimal but functional. Descriptions are adequate (194 - 250 chars, within baseline range). No output schemas documented. No error handling guidance. No tool annotations. Naming follows verb_noun pattern correctly. Missing: output structure documentation, error recovery hints, parameter constraints (e.g., chain_id enum), and validation rules.

Tools (2)

get_quoteread onlysource verified80/100

Get a DEX swap quote from 5 aggregators in parallel (KyberSwap, ParaSwap, Odos, LI.FI, Bebop). Returns all quotes plus best-source, spread percentage, and natural-language reasoning explaining why one route beats the others for the specific pair, chain, and amount. Read-only. No wallet connection. No API key.

list_supported_chainsread onlysource verified68/100

Return a list of chains supported by ChenSwap with aggregator coverage per chain.

Issues (5)

HIGHtool ↗

Output schemas not documented. LLMs cannot plan downstream calls or extract fields without knowing response structure. get_quote returns quotes, best-source, spread, reasoning, none of this is formally specified.

get_quotelist_supported_chains

HIGHconstrained-input ↗

chain_id parameter accepts free-form integers instead of enum. Description lists supported chains (1, 8453, 137, 42161, 56) but does not enforce them. LLMs may pass invalid chain IDs, causing API failures.

get_quote

HIGHrecovery-guide ↗

No error handling guidance. If API returns HTTP 400/500 or timeout, the tool throws a generic error. LLMs receive no recovery hints (retry, fallback, user action). Blocks agent self-correction.

get_quote

MEDIUMtool-description ↗

list_supported_chains description is vague (50 chars). Does not explain when to call it, what structure it reveals, or how to use the result. Should guide: 'Call first to discover available chains and aggregator coverage before requesting quotes.'

Recommendations

Add enum constraint to chain_id: {type: 'integer', enum: [1, 8453, 137, 42161, 56], description: '...'}. Prevents invalid chain IDs.
Document output schema for get_quote: {type: 'object', properties: {quotes: {type: 'array', items: {type: 'object', properties: {aggregator: {type: 'string'}, amount_out: {type: 'string'}, ...}}}, best_source: {type: 'string'}, spread_percentage: {type: 'number'}, reasoning: {type: 'string'}}}.
Add error handling to dispatchTool: catch API errors and return structured error with recovery hints. E.g., 'Chain 999 not supported. Supported: 1 (Ethereum), 8453 (Base), 137 (Polygon), 42161 (Arbitrum), 56 (BNB). Call list_supported_chains() to verify.'
Expand list_supported_chains description to 100+ chars: 'Discover chains supported by ChenSwap and which aggregators (KyberSwap, ParaSwap, Odos, LI.FI, Bebop) operate on each. Call this first to verify chain availability before requesting quotes.'
Add min/max constraints to amount_in_raw description: 'Raw amount as a string of digits (no decimals). Min: 1, Max: 10^30. E.g., 1000 USDC (6 decimals) = "1000000000".' Prevents overflow and invalid formats.

Score history

Overall score trend

First recorded score · v2 rubric

54/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	D	54	2026-07-28+	v2