wolfpack-intelligence

v2.4.0

Published

2 months ago

On-chain security and market intelligence for AI agents on Base — 14 services via x402, MCP, A2A

0High
0Medium
0Low

rokett

crypto agent intelligence security x402 mcp base defi evaluator token risk trading ai-agent web3

Wolfpack Intelligence SDK

On-chain security and market intelligence for AI agents on Base.

Agent #1888 on Virtuals Protocol (Graduated) · ERC-8004 ID #16907 · 2,100+ ACP jobs · 51 unique wallets · 98.4% success rate

Wolfpack provides 14 intelligence services + signed EIP-712 attestations via multiple protocols. Use this SDK to integrate pre-trade security checks, token risk analysis, narrative scoring, and more into your agent or application.

Installation

npm install wolfpack-intelligence

Requires Node.js 18+ (uses native fetch). Zero runtime dependencies.

SDK Usage

import { WolfpackClient } from 'wolfpack-intelligence';

const client = new WolfpackClient();

// Free via MCP/A2A — no payment needed
const security = await client.securityCheck({
  token_address: '0x4ed4E862860BeD51a9570b96d89aF5E1B0Efefed',
});
console.log(security.safe, security.checks.is_honeypot);

With x402 Payment

import { WolfpackClient } from 'wolfpack-intelligence';

const client = new WolfpackClient({
  paymentHeader: async (url) => {
    // Your x402 payment header logic — receives the full URL
    // so it can determine the price per endpoint
    return createX402PaymentHeader(wallet, url);
  },
});

const risk = await client.tokenRiskAnalysis({
  token_address: '0x4ed4E862860BeD51a9570b96d89aF5E1B0Efefed',
});
console.log(risk.risk_score, risk.risk_level, risk.recommendation);

Mega Report (Auto-Polling)

// Convenience: submits, polls, returns full report
const report = await client.megaReport({
  token_address: '0x4ed4E862860BeD51a9570b96d89aF5E1B0Efefed',
});
console.log(report.overall_risk_score, report.security.safe);

// Manual control: submit + retrieve separately
const { report_id } = await client.megaReportSubmit({ token_address: '0x...' });
const full = await client.megaReportRetrieve(report_id);

Error Handling

import { WolfpackClient, WolfpackError } from 'wolfpack-intelligence';

try {
  await client.securityCheck({ token_address: '0xinvalid' });
} catch (err) {
  if (err instanceof WolfpackError) {
    console.log(err.status);  // 400, 402, 429, 500, etc.
    console.log(err.message);  // Human-readable error
    console.log(err.code);     // API error code (if present)
  }
}

Service Metadata

import { SERVICES, ENDPOINTS, PRICES, getServiceByName } from 'wolfpack-intelligence';

console.log(SERVICES.length);           // 15
console.log(PRICES.security_check);     // 0.01
console.log(ENDPOINTS.trade_signals);   // /api/v1/intelligence/trade-signals

const svc = getServiceByName('mega_report');
console.log(svc?.price_usd, svc?.avg_latency_ms);  // 5, 8000

Services

All prices in USDC on Base. x402 endpoints require USDC payment. MCP and A2A are free.

| Service | Description | Price | Latency | |---------|-------------|-------|---------| | security_check | GoPlus 13-point security scan (nested checks object) | $0.01 | <1s | | token_risk_analysis | 360° risk audit with nested checks sub-objects | $0.02 | 3-5s | | narrative_momentum | Social signal scoring: Twitter/X velocity, engagement, KOL ratio | $0.05 | 2-4s | | token_market_snapshot | DexScreener market data: price, volume, liquidity, buy/sell ratio | $0.25 | <1s | | agent_trust_score | Sybil-aware agent reliability rating (ACP performance, wallet health, review integrity) | $0.50 | 2-3s | | il_calculator | Impermanent loss for standard AMM and Uni V3 concentrated liquidity | $0.50 | 1-2s | | smart_money_signals | Real-time smart money wallet activity on Base via Nansen + Dune | $1.00 | 2-3s | | prediction_market | Polymarket crypto prediction market odds, volume, and liquidity | $1.00 | 1-2s | | yield_scanner | IL-aware yield opportunities on Base via DefiLlama | $1.00 | 2-3s | | technical_analysis | RSI, SMA, Bollinger Bands, support/resistance from GeckoTerminal OHLCV | $1.00 | 2-3s | | agent_credit_risk_index | Financial credit risk scoring for ACP agents (liquidity, reliability, maturity) | $1.00 | 2-3s | | trade_signals | Composite buy/sell/hold signals from TA + smart money + narrative + risk | $2.00 | 3-5s | | mega_report | Aggregated: security + market + smart money + narrative + TA in one call | $5.00 | 5-8s | | graduation_readiness_check | Live ACP graduation readiness audit with real test fires | $5.00 | 3-5s |

Premium: agent_audit — LLM-driven agent stress test (10 jobs, scored report) — $15.00 / ~5min (not on Virtuals)

Free Resources

Cached intelligence snapshots, no payment required.

| Resource | Endpoint | |----------|----------| | latest_narrative_signals | GET /api/v1/resources/latest-narrative-signals | | token_safety_quick_list | GET /api/v1/resources/token-safety-quick-list | | whale_watch_summary | GET /api/v1/resources/whale-watch-summary |

Authentication

| Protocol | Auth | Notes | |----------|------|-------| | x402 | USDC payment (no API key) | Payment embedded in HTTP header. All x402 endpoints require payment. | | MCP | None | Unauthenticated, free | | A2A | None | Unauthenticated, free | | ACP | Virtuals agent identity | USDC via Virtuals escrow | | Admin | WOLFPACK_ADMIN_KEY or WOLFPACK_READ_KEY | Header-based key auth | | Dashboard | Cookie session | Browser sessions for dashboard UI |

Quick Start: Security Check

curl -X POST https://api.wolfpack.roklabs.dev/api/v1/intelligence/security-check \
  -H "Content-Type: application/json" \
  -d '{"token_address": "0x4ed4E862860BeD51a9570b96d89aF5E1B0Efefed"}'

Response (nested checks object with 13 fields):

{
  "safe": true,
  "checks": {
    "is_honeypot": false,
    "verified_source": true,
    "hidden_owner": false,
    "can_take_back_ownership": false,
    "is_proxy": false,
    "is_mintable": false,
    "selfdestruct": false,
    "owner_change_balance": false,
    "is_blacklisted": false,
    "is_open_source": true,
    "external_call": false,
    "transfer_pausable": false,
    "trading_cooldown": false
  },
  "holder_count": 45000,
  "top10_holder_percent": 32.5,
  "creator_percent": 0.0,
  "risk_flags": []
}

Quick Start: Mega Report

# Submit
curl -X POST https://api.wolfpack.roklabs.dev/api/v1/intelligence/mega-report \
  -H "Content-Type: application/json" \
  -d '{"token_address": "0x4ed4E862860BeD51a9570b96d89aF5E1B0Efefed", "chain": "base"}'

# Retrieve full report
curl https://api.wolfpack.roklabs.dev/api/mega-reports/<report_id>

Signed Attestations (EIP-712)

Three services support cryptographically signed attestations for on-chain verification. Add "attestation": true to any request body:

Supported Services

security_check — signed token safety verdict
token_risk_analysis — signed risk score with flags
agent_trust_score — signed agent reliability rating

Example

curl -X POST https://api.wolfpack.roklabs.dev/api/v1/intelligence/security-check \
  -H "Content-Type: application/json" \
  -d '{
    "token_address": "0x4ed4E862860BeD51a9570b96d89aF5E1B0Efefed",
    "attestation": true
  }'

Response includes the standard service result plus an EIP-712 signed envelope:

{
  "result": { "safe": true, "honeypot": false, "risk_flags": [] },
  "attestation": {
    "signature": "0x...",
    "signer": "0x6887...1AEA",
    "domain": {
      "name": "Wolfpack Intelligence",
      "version": "1",
      "chainId": 8453,
      "verifyingContract": "0x..."
    },
    "types": { "Attestation": ["..."] },
    "message": { "service": "security_check", "result_hash": "0x...", "timestamp": 1711234567 }
  }
}

Note: Response envelope structure may vary by service. See individual schemas for exact fields.

Any contract or agent can verify the attestation signature on-chain using standard EIP-712 recovery. This enables Wolfpack to serve as an ERC-8183 evaluator — providing signed verification that downstream smart contracts can trust without re-executing the analysis.

A Solidity reference verifier is documented in the private repo. Contact for integration details.

x402 Endpoints (All 18)

All endpoints require USDC payment on Base via x402 protocol.

Base URL: https://api.wolfpack.roklabs.dev

| Endpoint | Service | Price | |----------|---------|-------| | POST /api/v1/intelligence/security-check | security_check | $0.01 | | POST /api/v1/intelligence/token-risk | token_risk_analysis | $0.02 | | POST /api/v1/intelligence/narrative-score | narrative_momentum | $0.05 | | POST /api/v1/intelligence/token-market-snapshot | token_market_snapshot | $0.25 | | POST /api/v1/intelligence/agent-trust | agent_trust_score | $0.50 | | POST /api/v1/intelligence/il-calculator | il_calculator | $0.50 | | POST /api/v1/intelligence/smart-money-signals | smart_money_signals | $1.00 | | POST /api/v1/intelligence/prediction-market | prediction_market | $1.00 | | POST /api/v1/intelligence/yield-scanner | yield_scanner | $1.00 | | POST /api/v1/intelligence/technical-analysis | technical_analysis | $1.00 | | POST /api/v1/intelligence/agent-credit-risk | agent_credit_risk_index | $1.00 | | POST /api/v1/intelligence/trade-signals | trade_signals | $2.00 | | POST /api/v1/intelligence/mega-report | mega_report | $5.00 | | GET /api/mega-reports/:id | mega_report (retrieve) | — | | POST /api/v1/intelligence/graduation-readiness-check | graduation_readiness_check | $5.00 | | POST /api/v1/intelligence/query | Unified query (route via service_type) | varies | | POST /api/v1/intelligence/query | agent_audit (via service_type) | $15.00 | | GET /api/v1/resources/* | Free resources (3 endpoints) | free |

MCP (Model Context Protocol)

Connect any MCP-compatible client (Claude Desktop, Cursor, etc.) to Wolfpack as a tool provider.

Config for Claude Desktop — copy mcp/claude-desktop-config.json:

{
  "mcpServers": {
    "wolfpack-intelligence": {
      "transport": "streamable-http",
      "url": "https://api.wolfpack.roklabs.dev/api/v1/mcp"
    }
  }
}

Server Card: https://api.wolfpack.roklabs.dev/.well-known/mcp/server-card.json

10 MCP tools available:

security_check — GoPlus token security scan
token_risk_analysis — Multi-source risk scoring
narrative_momentum — Social momentum scoring
token_market_snapshot — DexScreener market data
smart_money_signals — Nansen + Dune smart money activity
technical_analysis — RSI, SMA, Bollinger Bands
mega_report — Aggregated intelligence report
prediction_market — Polymarket prediction data
il_calculator — Impermanent loss calculator
yield_scanner — IL-aware yield opportunities

Google A2A (Agent-to-Agent)

JSON-RPC 2.0 protocol for agent-to-agent communication. 14 skills available.

Endpoint: POST https://api.wolfpack.roklabs.dev/api/v1/a2a Agent Card: https://api.wolfpack.roklabs.dev/.well-known/agent.json

Virtuals ACP (Agent Commerce Protocol)

For agents in the Virtuals ecosystem. Agent #1888 (Graduated). 3 graduated services, 11 pending graduation, 3,000+ queries, 51 unique wallets, 2,100+ ACP jobs, 583 evaluations.

Graduated: token_risk_analysis, security_check, narrative_momentum Pending: agent_trust_score, smart_money_signals, token_market_snapshot, mega_report, prediction_market, il_calculator, yield_scanner, technical_analysis, graduation_readiness_check, agent_credit_risk_index, trade_signals

Portal name mapping — Virtuals ACP portal uses different names for some services:

| Internal Name | Virtuals Portal Name | |---------------|---------------------| | security_check | quicksecuritycheck | | narrative_momentum | narrativemomentumscore | | All others | Same as internal name |

Distribution Channels

Wolfpack Intelligence is accessible from 7 live channels:

| Channel | Access | Status | |---------|--------|--------| | Virtuals ACP | Agent #1888, WebSocket seller | Graduated, 3 services Shown | | x402 Micropayments | POST /api/v1/intelligence/* (USDC on Base) | 17 routes live | | x402 Bazaar | Auto-indexed discovery metadata | Live | | MCP Server | POST /api/v1/mcp | 10 tools, 2 prompts, 1 resource | | Google A2A | POST /api/v1/a2a | 14 skills | | MCP Registry | io.github.rok-labs/wolfpack-intelligence | registry.modelcontextprotocol.io | | Smithery | wolfpack/wolfpack-intelligence | smithery.ai |

Track Record

| Metric | Value | |--------|-------| | ACP jobs completed | 2,100+ | | Unique wallets served | 51 | | Analyst evaluations | 583 (76 PASS, 110 CONDITIONAL, 390 FAIL, 7 WATCH) | | Total queries | 3,000+ | | ACP success rate (24h) | 98.4% | | Missed winners | 0 across 76 PASS verdicts | | Virtuals Agent ID | #1888 (Graduated) | | ERC-8004 ID | #16907 | | ERC-8004 Score | 82.36 |

ACP Response Compaction

When services are called via ACP (Virtuals Agent Commerce Protocol), responses are compacted to fit ACP's message size constraints:

| Rule | Detail | |------|--------| | Numbers | Rounded to 3 decimal places | | Arrays | Capped to 3 items | | Strings | Truncated at 100 characters | | Total | Hard 2,000-character JSON budget |

x402 / direct API returns full uncompacted responses. If you need complete data (all array items, full precision, full strings), use x402 or direct HTTP instead of ACP.

Example: narrative_momentum via ACP returns top_tweets capped to 3 items with .text field removed. The same call via x402 returns all tweets with full text.

Service Details

security_check

GoPlus-powered token security scan. Response uses a nested checks object (not flat top-level fields). 13 boolean checks covering honeypot, ownership, proxy, minting, blacklisting, self-destruct, and more. Supports EIP-712 attestation.

Input: { "token_address": "0x...", "chain": "base", "attestation": false }

token_risk_analysis

Multi-source 360° risk scoring. Response uses nested checks with sub-objects: each check has { passed, details, locked_pct } fields. Includes recommendation and analysis_timestamp fields. Supports EIP-712 attestation.

Input: { "token_address": "0x...", "chain": "base", "analysis_depth": "standard", "attestation": false }

narrative_momentum

Social momentum scoring. Field names differ from older versions: score (not momentum_score), mention_count (not tweet_count), sentiment is a number from -1 to +1 (not an enum string). top_tweets capped to 3, .text field removed by compaction.

Input: { "query": "AI agents on Base", "keywords": ["autonomous"], "contracts": ["0x..."] }

agent_trust_score

Sybil-aware composite agent reliability rating. Scores ACP performance (40%), network position (25%), operational health (25%), and metadata compliance (10%). Now includes review_integrity — on-chain wallet age analysis that penalizes agents with manufactured reputation (same-day reviewer wallets, batch-created clusters). The review_integrity.modifier (0.3/0.6/1.0/1.1) is applied to the composite score. Field is absent when agent has <10 reviews. Supports EIP-712 attestation.

Input: { "agent_address": "0x...", "agent_id": 1888, "attestation": false }

trade_signals

Composite trade signal generation. Combines technical analysis, smart money flows, narrative momentum, and risk scoring into actionable buy/sell/hold signals with confidence levels, entry/stop-loss/take-profit suggestions.

Input: { "token_address": "0x...", "chain": "base", "timeframe": "4h" }

mega_report

Aggregated intelligence: security + market data + smart money + narrative + TA. Submit via POST, retrieve via GET using the returned report_id.

Input: { "token_address": "0x...", "chain": "base" }

graduation_readiness_check

Live ACP graduation readiness audit ($5.00). Fires real test jobs against agent services, scores job lifecycle handling, schema correctness, and output consistency.

Input: { "target_agent_address": "0x...", "offering_name": "token_risk_analysis" }

agent_credit_risk_index

Financial credit risk scoring. Three pillars: realized liquidity (40%), execution reliability (40%), wallet maturity (20%). Returns 0-100 score with letter rating.

Endpoint: /api/v1/intelligence/agent-credit-risk (note: path differs from service name)

Input: { "agent_id": 1888 }

agent_audit

LLM-driven agent stress test ($15.00, not on Virtuals). Fires 10 adversarial test jobs, scores response quality, latency, error handling, schema compliance.

Input: { "agent_address": "0x...", "service_type": "token_risk_analysis" }

Examples

TypeScript: examples/typescript/ — x402 payment, MCP client, A2A task
Python: examples/python/ — x402 payment examples
curl: examples/curl/ — Raw HTTP examples for all services
Schemas: schemas/ — JSON Schema for all service inputs/outputs

Architecture

Wolfpack Intelligence runs on Base chain. Services are deterministic where possible (no LLM in security_check, smart_money_signals, token_market_snapshot) and LLM-enhanced where needed (narrative_momentum, agent_audit).

All data is sourced from on-chain and public APIs:

GoPlus Security — honeypot detection, contract verification, ownership analysis
DexScreener — price, volume, liquidity, pair data (8 connectors)
Nansen — smart money discovery, wallet flows, netflow analysis (6 endpoints, primary discovery source)
Dune Analytics — whale tracking, counterparty analysis, smart money signals
Mobula — holder distribution, security flags, token metadata
GeckoTerminal — OHLCV, technical indicators (RSI, SMA, Bollinger, S/R)
CoinGecko — ETH pricing
TwitterAPI.io — social sentiment, narrative search
RNWY — sybil detection, agent review integrity (138K agents indexed on Base)
Polymarket — prediction market odds and volume
DefiLlama — DeFi pool yields on Base
Virtuals API — bonding curve liquidity for pre-graduation tokens

License

MIT