Loading ToolBench
Cross-chain DeFi intelligence for AI agents. Risk-scored yields, profit simulation, whale tracking across 86 chains and 6,500+ pools. The Bloomberg Terminal for AI.
ProfitSpot MCP demonstrates solid definition quality with well-described tools and complete input schemas. All 7 tools have descriptions exceeding 100 characters with clear use cases and tier information (FREE vs PRO). Input parameters are fully typed with descriptions, defaults, and constraints. However, output schemas are not formally documented in the source, descriptions mention what is returned (e.g., 'risk grade with full breakdown', 'historical APY trends') but no structured schema definitions are visible. Error handling is tier-gated (pro_gate, rate_gate) but lacks recovery guidance in responses. Tool naming follows verb_noun convention (discover_, analyze_, calculate_, simulate_, risk_, track_, defi_). Some tools conflate multiple responsibilities (e.g., discover_yields returns pool name, protocol, chain, APY, TVL, risk grade, prediction, and stablecoin flag in one response, good for the use case but information-dense). No security concerns detected; API keys are injected via environment (PROFITSPOT_API_KEY) and not exposed as parameters. Rate limiting is enforced. Overall: strong descriptive documentation, complete schemas, but missing formal output schema documentation and error recovery guidance.
Deep analysis of a specific DeFi pool. Returns: full pool data, historical APY trends (7d/30d/90d), TVL history, risk grade with full breakdown (why it got that grade), IL estimate scenarios for LP pools, and projected returns at current APY for $1K/$10K/$100K over 30/90/365 days. PRO ONLY — requires PROFITSPOT_API_KEY.
Calculate exact impermanent loss for a 50/50 LP pair. Given price change percentages for both tokens, calculates exact IL in USD and percentage. Also returns: value if held vs value in LP, and net gain/loss including trading fees at various APY levels (5%, 10%, 20%, 50%). ALWAYS FREE — pure math, no API call needed.
Big-picture DeFi dashboard. Returns: total DeFi TVL, TVL by top 10 chains, top 10 protocols by TVL, pool counts, and average yields. FREE tier: basic stats (TVL, chains, protocols). PRO tier: + average yields by risk grade, stablecoin market cap, hot opportunities (new pools with high APY + good risk grade).
Discover top DeFi yield opportunities across 86 chains and 6,500+ pools. Filter by chain, minimum TVL, minimum APY, and maximum risk grade (A-F). Returns sorted list with pool name, protocol, chain, APY, TVL, risk grade, prediction (stable/up/down), and stablecoin flag. FREE tier: limited to 10 results, no risk filtering. PRO tier: up to 50 results with full risk grade filtering.
Output schemas are not formally documented. Tool descriptions mention what is returned (e.g., 'historical APY trends (7d/30d/90d), TVL history, risk grade with full breakdown') but no structured JSON Schema or TypedDict is visible in the source code for return types. LLMs cannot reliably extract or chain results without formal schema documentation.
Error responses lack recovery guidance. When pro_gate() or rate_gate() return an error, the response tells the LLM the call failed but not what to do next (e.g., 'Set PROFITSPOT_API_KEY to upgrade to Pro tier' or 'Retry after 60 seconds'). This forces the agent to reason about recovery without hints.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | B | 74 | 2026-07-28+ | v2 |
Get risk grade (A-F) for a specific pool or entire protocol. Returns overall grade, breakdown scores for: TVL stability, APY sustainability + volatility, protocol reputation + age, IL exposure, pair stability, and security/audit proxy. Each sub-score explained in plain English. For protocols: scores the top pools and returns an aggregate grade. PRO ONLY — requires PROFITSPOT_API_KEY.
Project returns for a specific pool using Monte Carlo simulation. Uses Ornstein-Uhlenbeck APY model that accounts for: APY mean reversion, APY crash events, impermanent loss via geometric Brownian motion, pool failure probability, and gas costs. Runs 1,000 simulations. Returns optimistic/base/bearish scenarios with dollar amounts, percentiles, probability of profit, and risk factors. PRO ONLY — requires PROFITSPOT_API_KEY.
Detect large capital movements (whale activity) across DeFi pools. Finds pools where TVL changed by >$500K since last snapshot. Reports: pool, direction (inflow/outflow), amount, percentage change, and whether this correlates with APY changes. First call establishes a baseline; subsequent calls detect actual changes. Also uses APY-change heuristics as a proxy for whale activity. PRO ONLY — requires PROFITSPOT_API_KEY.
discover_yields enforces tier-based limits (FREE: 10 results, no risk filtering; PRO: 50 results with risk filtering) but returns a 'tier_note' string in the response only when risk filter is disabled, making the note advisory rather than assertive. The tool description mentions the limits but downstream LLM reasoning may not catch the constraint if the note is buried in response data.
Output field naming not explicitly documented. E.g., discover_yields returns 'pools' but the description does not formally specify the structure or field names (pool_name, protocol, chain, apy, tvl, risk_grade, prediction, is_stablecoin). LLMs infer field names from prose, mismatches force parsing failures.
PRO-only tools (analyze_pool, simulate_profit, risk_score, track_whales) accept parameters but gate execution. A clearer pattern would be to document tier requirements in the parameter schema (marking certain params as requiring PRO) rather than hiding the gate in tool implementation. This prevents LLM confusion about when a tool is actually callable.