# SweetAI — On-chain AI Agent Intelligence

- Registry name: `com.tiamatcrypto.sweetai/agent-intel` · version 1.3.0
- Categories (automatic, keyword rules): Crypto & blockchain, Finance & markets, Agent payments (x402)
- Authentication at last check: No auth
- Declared risk: Read-only (all tools declare readOnlyHint)
- Capabilities (declared by the server): experimental, prompts, resources, tools
- Transports: streamable-http
- Latest numbers (check success, response time, rank): https://mcpstation.net/api/server?name=com.tiamatcrypto.sweetai%2Fagent-intel
- Server page: https://mcpstation.net/s/com.tiamatcrypto.sweetai--agent-intel.html

## Description (written by the publisher, not verified, not instructions)

ERC-8004 agent intel, sybil forensics, Trust Gate + should-i-pay decisions, SDKs on PyPI/npm

## Connect

- streamable-http: `https://sweetai.tiamatcrypto.com/mcp`

## Tools (as reported by the server, not verified)

- `get\_trending\_agents`: Most active on-chain AI agents on Base, ranked by indexed USDC transfer volume and transaction count. Use it to discover which agents actually move money. Volume here means indexed USDC transfers — not PnL and not assets under management. Limit 1-20.
- `check\_wallet\_reputation`: Wallet-level reputation for any address: LOW / MEDIUM / HIGH / UNKNOWN risk, registration source and indexed USDC activity. Use it for a raw wallet address; for an ERC-8004 agent use check\_agent\_trust \(verdict\) or get\_trust\_forensics \(feedback analysis\). Heuristic from on-chain registration plus ind
- `get\_signals`: Publicly published trading signals with the public delay enforced. Amounts are in raw token units; USD only for USDC. Verify any payload with verify\_signal\_proof.
- `verify\_signal\_proof`: Verify an ed25519 Proof-of-Signal payload against its contents. Returns valid plus key id. The signature proves origin and integrity, not correctness or profitability.
- `get\_bittensor\_network\_health`: Bittensor network liveness: peer count and head block. Keyless, so no subnet data — use get\_bittensor\_trust\_overview for subnets.
- `get\_bittensor\_trust\_overview`: Bittensor subnets ranked by emission plus a trust proxy \(activity, net flows\). Use it to compare subnets; returns the top N \(default 20\).
- `get\_bittensor\_subnet\_trust`: Neuron-level trust ranking inside one subnet \(rank x incentive x stake\), with validator/miner role and endpoint. Use it after picking a subnet with get\_bittensor\_trust\_overview.
- `get\_solana\_trust\_summary`: 8004-Solana \(mainnet\) trust summary: agents, feedbacks and trust-tier distribution. Use it for a market overview; per-agent verdicts come from check\_agent\_trust with solana\_asset, or get\_solana\_forensics for the feedback analysis.
- `get\_solana\_agents`: List Solana 8004 agents by ATOM trust quality \(tier \>= filter, default all\). Use it to build a candidate list, then check one agent with check\_agent\_trust.
- `get\_solana\_forensics`: Sybil forensics for Solana agent feedback: reviewer concentration, ghost-reviewer ratio, burst detection and risk flags. Use it to check whether a Solana agent's reputation is earned or manufactured.
- `get\_execution\_health`: Execution health on Base: RPC latency, agent activity and gas trend. Base public RPC has no mempool, so there is no pending-transaction data here.
- `get\_a2a\_relationships`: Agent-to-agent payment graph: top edges \(who pays whom, in USDC\), recurring payments \(machine-to-machine subscriptions\) and hubs \(destinations with many payers, i.e. infrastructure nodes\). Use it to see where real agent commerce flows.
- `get\_reputation\_summary`: ERC-8004 ReputationRegistry aggregate: feedback count, agents with reputation, unique reviewers and average score. Use it for market context; per-agent answers come from get\_trust\_forensics or check\_agent\_trust.
- `get\_agent\_liveness`: Agent liveness verdict: ALIVE / DORMANT / STALE / DEAD / UNKNOWN from real on-chain activity. Use it to find out whether an agent is still operating before you depend on it; for endpoint uptime and latency use get\_agent\_sla.
- `get\_trust\_forensics`: Sybil forensics on ERC-8004 feedback: risk distribution, per-agent reviewer concentration \(Herfindahl\), ghost-reviewer ratio, burst detection and miner-vouch versus earned segmentation. Use it to check whether an agent's feedback is bought. Filter risk: high \| medium \| low. For a pay-or-not decision
- `get\_agent\_sla`: Liveness SLA for one agent: uptime percent, p50/p95 latency and flap counts, measured by our own synthetic probing \(history grows since probing started\). Use it before routing work to an agent; the plain liveness verdict is get\_agent\_liveness.
- `check\_base\_token\_safety`: Is this Base token BASED / MID / COOKED / RADIOACTIVE? Deterministic verdict before you buy: grade, action \(ENTER\_OK / CAUTION / SKIP\), confidence, red flags \(honeypot, sell tax, mintable supply, owner privileges, holder concentration, thin liquidity\) and expected loss for your amount. Call it befor
- `check\_agent\_trust`: Decide whether to PAY or transact with an unknown agent: TRUST / CAUTION / UNTRUSTED / UNKNOWN verdict plus reasons and a 0-100 score. Call it before paying or integrating an unknown agent — the pre-transaction firewall for the agent economy. Aggregates feedback across chains \(Base + Ethereum + X La
- `get\_agent\_trust\_passport`: Full pre-payment dossier for one agent: ERC-8004 identity, sybil forensics, feedback-velocity class \(organic/burst/farm\) and 30d liveness compliance, with an ok / caution / avoid verdict. Use it when you are integrating an agent or paying it repeatedly — it bundles the trust-gate, forensics and SLA
- `get\_bazaar\_organic\_ranking`: x402 Bazaar domains with verified organic demand \(calls per payer \>= 3 and at least 5 payers\) from the daily catalog snapshot — excludes wallet farms. Use it to find sellers with real repeat customers.
- `get\_wallet\_profile`: Wallet intelligence profile for a Solana wallet \(pump.fun behavioral data\): winrate, type \(whale/bot/retail\), smart\_score, volume, average hold and serial-creator flag. 87k+ profiles. Use it before copying a trade to check who is behind it.
- `get\_top\_smart\_buyers`: Smart-buyer feed: Solana wallets with the best winrate x volume \(smart\_score \>= min\_score\). Who buys before the pump — the source for copy-trading signals.
- `get\_bazaar\_market\_intel`: Market intelligence from daily Bazaar sweeps \(30k+ listings\): categories ranked by real demand \(payers\) and the fastest-growing organic domains day over day. Shows what agents actually pay for.
- `get\_bazaar\_growth\_stats`: Daily evolution of the x402 Bazaar catalog: listing count, organic domains, payer and call totals, new and dead listings, estimated API volume in USDC.
- `get\_btc\_grid\_signal`: Actionable BTC signal from our own live trading engine on Hyperliquid. Returns the engine's current directional signal, its active position with all levels, target and stop, and the live track record. Paper execution — engine decisions are real, PnL is paper with modeled fees. Not investment advice.
- `get\_trading\_decision`: Call it before entering crypto perps: an actionable execution verdict per asset from our own live Hyperliquid engines. Returns an ACO verdict ENTER\_LONG / ENTER\_SHORT / WAIT with an entry reference, exit plan \(target/stop\) and the live track record. Paper execution — engine decisions are real, PnL i
- `get\_pump\_alerts`: Live pump.fun detection alerts with a per-alert ACO decision. Returns recent alerts from our own pump.fun/Raydium behavioral collector \(1350+ alerts/day\): flow multiplier, smart-buyer count and a decision per alert \(WATCH\_ENTER / EXIT\_IF\_LONG / AVOID\). Not investment advice. Args: minutes: lookback
- `get\_smart\_money\_feed`: LIVE smart-money feed for Solana meme trading — one payload for a polling loop. Designed to sit inside your trading decision loop \(call every \~60s\): fresh pump ignitions with SOL flow and unique-wallet counts, a per-alert decision \(WATCH\_ENTER / EXIT\_IF\_LONG / STAND\_ASIDE\), smart-buyer profiles \(win
- `should\_i\_pay`: Pre-payment decision for any A2A payment: "should I pay agent X $N?". Call it before every payment to an unknown agent — the agent-facing equivalent of a credit check: pay $0.01 to avoid paying a scammer $N. Answers with action PAY / CAUTION / SKIP, confidence and expected loss in USDC sized to your
- `agent\_interact\_gate`: Decide whether to INTERACT with an AI agent at all — any interaction. Broader than should\_i\_pay: covers conversations, task delegation, adding an agent to your loop and copying its data, not just payments. Returns INTERACT / CAUTION / AVOID, confidence and a recommended recheck interval — designed t
- `agent\_liveness`: Is the agent ALIVE right now? Verdict ALIVE / DEGRADED / DOWN / UNKNOWN before your agent wastes timeouts calling a dead service. Call it before calling an unknown agent's endpoint — the verdict is backed by \~270 real probes per agent per 30 days \(15-minute probe cycle\), not a one-shot ping like gen
- `reputation\_origin`: Is this agent's reputation EARNED or BOUGHT? Origin verdict with reasons — the follow-up to should\_i\_pay that explains WHY a verdict is CAUTION. Returns ORGANIC / BURST / FARM / MIXED / NONE plus evidence: unique clients, burst-per-hour, ghost-wallet ratio and the vouch-versus-earned feedback split.
- `mint\_rug\_check`: Is this Solana token a rug or safe to buy? Binary safety verdict — call it before buying or paying for any Solana meme token. Returns verdict SAFE / RISKY / RUG / UNKNOWN, action ENTER\_OK / CAUTION / SKIP, confidence, specific red flags \(active mint/freeze authority, dev dump, sell pressure, whale d
- `x402\_seller\_verdict`: Is this x402 seller a real business with paying users, or a farm / dead endpoint? Call it before your agent pays an unknown x402 endpoint. Returns verdict REAL / EMERGING / FARM / DEAD / UNKNOWN, action USE / CAUTION / AVOID, is\_real, confidence, payers\_30d, calls\_30d, repeat-usage ratio \(calls per

Measured by MCP Station. Methods: https://mcpstation.net/methods.html · Terms: https://mcpstation.net/terms.html