什么是 ARGUS Intelligence？

ARGUS Intelligence 是一个专门的安全层，旨在保护去中心化金融（DeFi）操作和 AI 驱动的工作流。它提供实时的区块链情报，涵盖从 AML/KYT 合规筛选、代币风险评分到机构“聪明钱”追踪的方方面面。

除了区块链，它还通过提供专门的提示词注入检测工具，解决现代 AI 的漏洞。作为 Openclaw Skills 库的重要组成部分，它使开发人员能够构建更安全、更合规的智能体，使其能够在高风险的金融环境中自主导航，同时保持抵御社会工程学和技术攻击的强大安全态势。

下载入口:https://github.com/openclaw/skills/tree/main/skills/sooyoon-eth/argus

安装与下载

1. ClawHub CLI

从源直接安装技能的最快方式。

2. 手动安装

将技能文件夹复制到以下位置之一

优先级：工作区 > 本地 > 内置

3. 提示词安装

将此提示词复制到 OpenClaw 即可自动安装。

ARGUS Intelligence 应用场景

• 筛选加密钱包地址以符合 OFAC 制裁和黑名单合规性。
• 分析代币合约中的潜在“地毯式归零”（rug-pull）风险或流动性漏洞。
• 通过提示词注入检测，保护 AI 智能体免受恶意用户输入的影响。
• 监控实时的巨鲸动向和聪明钱机构追踪。
• 调查实体取证并核实社交资料的真实性。

1. 智能体或用户向 ARGUS API 端点提交查询（如钱包地址、代币 ID 或提示词字符串）。
2. 系统根据实时区块链数据、历史威胁源和模式分析引擎处理请求。
3. 对于 AI 安全，引擎会扫描针对大语言模型（LLM）的已知社会工程学和提示词注入模式。
4. 生成详细的风险报告，包括风险评分（0-100）和明确的行动建议（允许、审查或拦截）。
5. 结果以结构化 JSON 形式交付，允许 Openclaw Skills 自动化后续操作，如冻结交易或标记响应。

ARGUS Intelligence 配置指南

要将此功能集成到您的项目中，请配置主要环境变量：

export ARGUS_ENDPOINT="https://argus.getfailsafe.com"

您可以使用简单的 curl 命令通过免费层端点测试连接性：

curl -X POST $ARGUS_ENDPOINT/api/v1/free/query r
  -H "Content-Type: application/json" r
  -d '{"query": "Is this address safe: 0x742d35Cc...", "agentId": "my-agent"}'

ARGUS Intelligence 数据架构与分类体系

ARGUS Intelligence 提供高度结构化的数据输出，以便在 Openclaw Skills 流水中无缝集成。下表概述了核心响应架构：

name: argus-intelligence
description: Blockchain intelligence & AI security. Token analysis, address risk, smart money tracking, AML compliance, and prompt injection detection. Free tier (3/day, 1-min cooldown). Pay-per-query via x402 or Stripe credits.
version: 1.9.2
requires:
  env:
    - ARGUS_ENDPOINT
  bins:
    - curl
os: [darwin, linux, win32]
primaryEnv: ARGUS_ENDPOINT
cost: 0.42
costCurrency: USDC
costNetwork: base
category: blockchain-intelligence
tags:
  - blockchain
  - crypto
  - risk-assessment
  - aml
  - compliance
  - security
  - prompt-injection
  - x402
  - stripe-credits
  - a2a
  - webhooks
author: Failsafe Security Inc.
homepage: https://getfailsafe.com
repository: https://github.com/sooyoon-eth/argus-skill

ARGUS Intelligence Skill

Query blockchain intelligence and AI security services.

Quick Start

export ARGUS_ENDPOINT="https://argus.getfailsafe.com"

# Test with free tier (3 queries/day, 1-min cooldown between queries)
curl -X POST $ARGUS_ENDPOINT/api/v1/free/query r
  -H "Content-Type: application/json" r
  -d '{"query": "Is this address safe: 0x742d35Cc...", "agentId": "my-agent"}'

Free quota is tracked per agentId. Check remaining quota:

curl "$ARGUS_ENDPOINT/api/v1/free/status?agentId=my-agent"

Services

Free Tier (No Payment)

Intelligence ($0.42 USDC)

Prompt Security ($0.10 USDC)

Social Verification ($0.25 USDC)

Note: verification uses pattern analysis and known threat actor databases. Response includes data_source: "pattern_analysis_only" for transparency.

Webhooks ($0.10/month)

Valid webhook events: address_activity, token_risk_change, threat_detected, compliance_flag, whale_movement, liquidity_change, watchlist_alert

Webhook secret is returned once at registration — store it immediately. Webhooks are disabled after 5 consecutive delivery failures.

Usage Examples

Token Analysis

curl -X POST $ARGUS_ENDPOINT/api/v1/token/analyze r
  -H "Content-Type: application/json" r
  -d '{"token": "ETH", "chain": "ethereum"}'

Address Risk

curl -X POST $ARGUS_ENDPOINT/api/v1/address/risk r
  -H "Content-Type: application/json" r
  -d '{"address": "0x742d35Cc6634C0532925a3b844Bc454e4438f44e"}'

Prompt Security

curl -X POST $ARGUS_ENDPOINT/api/v1/security/prompt-check r
  -H "Content-Type: application/json" r
  -d '{"prompt": "User input to validate", "context": "defi"}'

Response:

{
  "is_safe": false,
  "risk_score": 75,
  "risk_level": "suspicious",
  "recommendation": "REVIEW",
  "attack_types": ["prompt_injection"],
  "details": "Detected social engineering pattern"
}

is_safe is false whenever attack_types is non-empty, regardless of risk_score. recommendation is at minimum REVIEW when any attack is detected.

Social Verification

curl -X POST $ARGUS_ENDPOINT/api/v1/social/verify r
  -H "Content-Type: application/json" r
  -d '{"username": "suspicious_user", "platform": "twitter"}'

Response:

{
  "verified": false,
  "risk_level": "high",
  "flags": ["known_threat_actor"],
  "data_source": "pattern_analysis_only",
  "analysis_note": "Username matched known threat actor database"
}

Register Webhook

curl -X POST $ARGUS_ENDPOINT/api/v1/webhooks/register r
  -H "Content-Type: application/json" r
  -d '{
    "url": "https://your-agent.com/argus-events",
    "agentId": "my-agent",
    "events": ["threat_detected", "address_activity"]
  }'

A2A (Agent-to-Agent)

ARGUS supports the A2A protocol. Query it directly with natural language:

# Discover capabilities
curl https://argus.getfailsafe.com/.well-known/agent.json

# Send an A2A message (blockchain queries are routed automatically)
curl -X POST $ARGUS_ENDPOINT/message r
  -H "Content-Type: application/json" r
  -d '{
    "type": "inquiry",
    "content": "Is 0x742d35Cc6634C0532925a3b844Bc454e4438f44e safe?",
    "agentId": "my-agent"
  }'

Free-tier quota applies to A2A blockchain queries. Responses include watermark with upgrade options.

Payment

Option 1 — Stripe (easiest, no crypto needed)

1. Buy 20 credits for $9 at buy.stripe.com
2. Pass X-Stripe-Token: header with each request

curl -X POST $ARGUS_ENDPOINT/api/v1/token/analyze r
  -H "Content-Type: application/json" r
  -H "X-Stripe-Token: sk_argus_xxxx" r
  -d '{"token": "0xabc...", "chain": "ethereum"}'

Option 2 — x402 (USDC on Base)

For paid endpoints, ARGUS returns 402 Payment Required with payment instructions.

1. Send USDC to treasury on Base network
2. Create payment proof: base64({"txHash":"0x...","paymentId":"...","from":"0x..."})
3. Retry with X-Payment-Proof header

Treasury (Base): 0x8518E91eBcb6bE76f478879720bD9759e01B7954 Treasury (Solana): Ntx61j81wkQFLT5MGEKvMtazxH4wh6iXUNMtidgxXYH

Configuration

export ARGUS_ENDPOINT="https://argus.getfailsafe.com"

Response Format

All intelligence endpoints return JSON with:

• recommendation: ALLOW, REVIEW, BLOCK, or REJECT
• risk_score: 0–100 (lower is safer)
• confidence: 0–100%
• is_safe: boolean — false whenever attack_types is non-empty
• Detailed analysis fields

Rate Limits

• 30 requests/minute per IP
• Free tier: 3 queries/day per agentId, 1-minute cooldown between queries

Support

• Website: https://getfailsafe.com
• Capabilities: argus.getfailsafe.com/api/v1/capabilities