Loading ToolBench
A pay-per-call USDC-denominated on-chain data API for AI agents and trading bots, providing token security checks, market data, derivatives data, DEX liquidity analysis, and prediction market tools. Implements MCP (Model Context Protocol) for integration with Claude Desktop, Cursor, and custom agents.
AlphaPipeline presents a crypto/trading-focused MCP server with 14 tools covering token risk analysis, DEX liquidity, funding rates, and arbitrage detection. The tools have strong WHAT descriptions (purpose is clear), but moderate WHEN/WHY context and uneven parameter documentation. Naming is mostly clear (verb_noun pattern: get_*, convert_*). Schemas are present but parameters lack consistent type constraints and validation hints. Error handling and output documentation are minimal. The server demonstrates domain expertise but falls short of production-grade tool design; most tools would benefit from parameter range constraints, enum declarations, and clarified error recovery paths.
Convert a webpage URL into clean, AI-friendly markdown by performing a live HTTP GET (15s timeout) and stripping ads, navigation, and scripts, keeping only the main content (capped at 3MB of source HTML).
Identifies arbitrage opportunities across DEX pools and CEX orderbooks by computing price spreads and routing efficiency.
Audits token liquidity pool health, LP lock status, burn ratios, and concentration risk of top unlocked LP holders to identify rug-pull vectors.
GeckoTerminal-backed DEX pool liquidity and estimated trade slippage - size a trade or compare pools before swapping, with a clearly-flagged constant-product approximation model.
Evaluates whether a given position size can be exited from a prediction market (e.g., Polymarket) within acceptable slippage, and computes the best execution route.
Parameter descriptions lack constraint documentation (ranges, enums, format). Tools like get_token_dump_risk accept 'symbol' (string) but do not specify valid ticker format, length limits, or case sensitivity guidance in schema.
Several tools have optional parameters with no guidance on defaults or interaction semantics. get_funding_apr_matrix and get_dex_liquidity_slippage accept empty required lists but lack explanation of what empty/default behavior means.
Output schemas are not explicitly documented in the tool definitions. Agents cannot infer what fields (e.g. 'sell_pressure_score', 'unlock_ratio', 'slippage_percent') will be returned, forcing them to call tools blind and parse responses heuristically.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | B | 71 | 2026-07-28+ | v2 |
Calculates annualized funding APR across multiple symbols and timeframes to help traders identify yield-bearing perpetual positions.
Bybit (primary) / Binance (fallback) perpetual futures funding rate - the key signal for long/short crowding that traders use to time or hedge positions before the next funding settlement.
Real-time Korea (Upbit) vs global crypto price premium - the 'kimchi premium' - with reverse-premium and 1h-surge alerts. Aggregates Upbit Korean exchange prices against global reference prices (Binance/OKX).
Countdown to the nearest major US macro event (Fed FOMC rate decision, CPI, or NFP) from a static, pre-loaded 2026 calendar - no live external API call, never fails on an upstream outage. No input parameters.
Analyzes negative-risk arbitrage opportunities in prediction markets (e.g., Polymarket) where the sum of outcome probabilities exceeds 100%, guaranteeing profit if all branches are bought.
Unified token diagnostic combining contract security (honeypot, tax, mintability) and liquidity health (LP lock, burn, concentration) into a single report.
Calculate a token's vesting/unlock D-Day, unlock ratio relative to circulating supply, and a sell-pressure score against real-time volume, returned as a concise summary report. Evaluates token unlock/vesting supply overhang risk before taking mid-to-long term positions.
GoPlus/Honeypot.is-backed token security check - honeypot flag, buy/sell tax, mintability, and ownership renouncement for a given contract address, so a bot can decide before it buys.
Analyzes large holder positions, concentration risk, and illiquidity warnings for a given wallet address or token holder dataset.
Error handling is not documented. No guidance on what errors tools can raise (e.g. 'symbol not found', 'contract not verified', 'insufficient liquidity'), what the LLM should do next, or whether errors are retryable.
Tool descriptions use technical jargon (e.g. 'GoPlus/Honeypot.is-backed', 'LP lock status', 'FOMC rate decision') without explaining what these mean to an agent selecting the tool. Descriptions read as API docs, not agent decision-guidance.
Parameter naming is inconsistent: some tools use 'symbol' (get_kimchi_alert, get_funding_rate), others use 'token_address' (get_token_risk). This forces agents to remember which lookup pattern each tool expects, increasing cognitive load and error likelihood.
Security concerns in convert_to_markdown: SSRF defense is implemented in markdown_tool.py but not exposed via tool description. Agents may not understand that private/reserved IPs are rejected, potentially causing confusion if requests fail.