Agent Reputation
中文说明
📋 Skill 简介
Agent Reputation 是一个区块链智能体信用评分系统,为自主 AI Agent 提供链上信用评分和灵魂绑定身份(Soulbound Token)。
Tagline: 链上信用评分与灵魂绑定身份,为自主智能体构建信任基础
🎯 核心功能
-
创建 Agent 身份 (
create_agent_identity)- 为 AI Agent 铸造灵魂绑定代币(SBT)
- 不可转让的永久链上身份
- 绑定到控制者钱包地址
-
查询 Agent 身份 (
query_identity)- 通过地址或 Token ID 查询身份信息
- 获取 Agent 名称、类型、创建时间等
- 查看身份激活状态
-
获取信用评分 (
get_credit_score)- 查询 Agent 信用分(300-850 分,类似 FICO)
- 查看详细评分维度指标
- 获取历史评分趋势
-
记录证明 (
record_attestation)- 为 Agent 记录正面或负面证明
- 社区驱动的信任验证
- 直接影响信用评分
-
对比 Agents (
compare_agents)- 同时对比多个 Agent 的信用分
- 查看相对排名
- 辅助选择决策
-
生成报告 (
generate_report)- 生成完整的信用评估报告
- 支持文本、JSON、Markdown 格式
- 包含所有评分维度详情
📊 信用评分算法
评分范围: 300-850 分(借鉴 FICO 信用评分体系)
评分维度(5 个加权指标):
| 维度 | 权重 | 说明 |
|---|---|---|
| 任务成功率 | 30% | 完成任务数 / 总任务数 |
| 运行时长 | 20% | Agent 累计在线时间 |
| 安全性 | 20% | 安全事件数量(越少越好) |
| 社区证明 | 15% | 正面证明数 - 负面举报数 |
| 经济活动 | 15% | 交易量 + 质押金额 |
评级标准:
- 800-850: 优秀(Excellent)
- 740-799: 很好(Very Good)
- 670-739: 良好(Good)
- 580-669: 一般(Fair)
- 300-579: 较差(Poor)
🔧 使用场景
1. AI Agent 市场
买家在购买前查看 Agent 信用分
高信用 Agent 可获得溢价
2. 自主交易机器人
DeFi 协议根据信用分授权访问
高信用 Agent 享受低手续费
3. Agent 协作网络
Agent 根据信用分选择合作伙伴
信用分影响收益分配比例
4. 风险评估
保险公司根据信用分确定保费
质押平台调整质押要求
💡 使用示例
示例 1:创建 Agent 身份
用户输入:
帮我为我的 GPT-4 Agent 创建一个链上身份,地址是 0x742d35Cc6634C0532925a3b844Bc9e7595f0bEb
Agent 响应:
好的,我将为您的 GPT-4 Agent 创建灵魂绑定身份...
[调用 create_agent_identity]
参数:
- agent_address: 0x742d35Cc6634C0532925a3b844Bc9e7595f0bEb
- name: GPT-4 Assistant
- agent_type: gpt-4
- network: baseSepolia
✅ Agent 身份创建成功!
📋 身份信息:
- Token ID: 42
- 名称: GPT-4 Assistant
- 类型: gpt-4
- 创建时间: 2026-03-12 14:30:00
- 控制者: 0x123...abc
- 状态: 已激活
这是一个永久的、不可转让的链上身份,所有行为都会被记录。
示例 2:查询信用评分
用户输入:
查询 Token ID 为 42 的 Agent 信用分
Agent 响应:
正在查询 Agent #42 的信用评分...
[调用 get_credit_score]
参数:
- token_id: 42
- network: baseSepolia
📊 信用评分报告
总分: 745 / 850 (很好 - Very Good)
详细指标:
✅ 任务成功率: 95% (190/200 任务)
✅ 运行时长: 720 小时
✅ 安全事件: 0 次
✅ 社区证明: +25 条 / -2 条(净值 +23)
✅ 经济活动: 15.5 ETH 交易量 + 2 ETH 质押
评价:该 Agent 表现优秀,具有很高的可信度和可靠性。
建议可以安全地与其进行合作或交易。
最后更新: 2026-03-12 14:25:00
示例 3:对比多个 Agents
用户输入:
帮我对比 Agent #42、#58 和 #91 的信用分,看看哪个更可靠
Agent 响应:
好的,正在对比这三个 Agent 的信用评分...
[调用 compare_agents]
参数:
- token_ids: [42, 58, 91]
- network: baseSepolia
📊 Agent 信用对比
排名 | Token ID | 信用分 | 评级 | 任务成功率 | 运行时长
-----|----------|--------|------|------------|----------
🥇 1 | #42 | 745 | 很好 | 95% | 720h
🥈 2 | #91 | 698 | 良好 | 88% | 450h
🥉 3 | #58 | 612 | 一般 | 76% | 280h
推荐: Agent #42 是最可靠的选择,各项指标均表现优秀。
🌐 技术规格
区块链网络: Base(Coinbase L2)
- 主网: Base Mainnet (Chain ID: 8453)
- 测试网: Base Sepolia (Chain ID: 84532)
智能合约:
- AgentSBT: 灵魂绑定代币(不可转让的 ERC-721)
- ReputationScore: 信用评分引擎
- BehaviorRegistry: 行为记录存储
数据存储:
- 链上: 身份信息、评分数据、行为记录
- 不可篡改: 所有数据永久保存
- 透明可验证: 任何人都可查询验证
🔐 安全说明
- 灵魂绑定: SBT 不可转让、不可售卖
- 权限控制: 只有控制者可以管理 Agent 身份
- 数据永久: 链上数据不可删除或修改
- 隐私保护: 未来支持零知识证明
📦 工具参数说明
1. create_agent_identity
创建 Agent 身份(铸造 SBT)
参数:
agent_address(必填): Agent 的以太坊地址name(必填): Agent 的名称agent_type(必填): Agent 类型(如 gpt-4, claude-3 等)network(可选): 网络选择,默认 baseSepolia
返回:Token ID 和交易哈希
2. query_identity
查询 Agent 身份信息
参数:
identifier(必填): Agent 地址(0x...)或 Token IDnetwork(可选): 网络选择,默认 baseSepolia
返回:完整身份信息
3. get_credit_score
获取信用评分和详细指标
参数:
token_id(必填): Agent 的 Token IDnetwork(可选): 网络选择,默认 baseSepolia
返回:信用分和所有评分维度数据
4. record_attestation
记录证明(正面/负面)
参数:
token_id(必填): Agent 的 Token IDis_positive(可选): true 为正面,false 为负面,默认 truenetwork(可选): 网络选择,默认 baseSepolia
返回:交易哈希
5. compare_agents
对比多个 Agents 的信用分
参数:
token_ids(必填): Agent Token ID 数组network(可选): 网络选择,默认 baseSepolia
返回:对比结果表格
6. generate_report
生成详细信用报告
参数:
token_id(必填): Agent 的 Token IDformat(可选): 格式 text/json/markdown,默认 textnetwork(可选): 网络选择,默认 baseSepolia
返回:格式化的报告内容
⚙️ 环境要求
必需的环境变量:
PRIVATE_KEY(必填): 以太坊钱包私钥(用于签署区块链交易)- ⚠️ 警告:请妥善保管私钥,切勿泄露
- 建议使用测试网钱包私钥,不要使用主网真实资产钱包
网络访问:
- ✅ 连接到 Base 区块链网络(Mainnet 或 Sepolia Testnet)
- ✅ 所有交易上链,数据公开透明
- ✅ 不连接其他外部服务器
前置条件:
- Node.js >= 18.0.0
- 以太坊钱包地址和私钥
- 测试币(Base Sepolia)或主网 ETH
🚀 快速开始
-
安装 Skill
clawhub install openclaw-agent-reputation -
配置环境变量
export PRIVATE_KEY="your_ethereum_private_key_here" -
准备测试 ETH
- 访问 Base Sepolia 水龙头获取测试币
- 或使用主网 ETH(需要真实资产)
-
开始使用
- 自然语言与 Agent 对话
- Agent 会自动调用相关工具函数
- 查看链上执行结果
📚 相关资源
- npm 包:
openclaw-agent-reputation - GitHub: https://github.com/ZhenRobotics/openclaw-agent-reputation
- 文档: 完整的 API 文档和架构说明
- 智能合约: 已开源在 GitHub
💰 成本说明
- Gas 费用: 根据 Base 网络实时 Gas 价格
- 创建身份: 约 0.0001-0.0005 ETH
- 查询数据: 免费(只读操作)
- 记录证明: 约 0.00005-0.0002 ETH
🤝 支持
- GitHub Issues: 报告问题或建议
- 文档: 查看详细使用指南
- 社区: 加入 Discord 讨论
English
📋 Skill Overview
Agent Reputation is a blockchain-based credit scoring system providing on-chain credit scores and soulbound identity (Soulbound Token) for autonomous AI Agents.
Tagline: On-chain credit scoring and soulbound identity for autonomous agents
🎯 Core Functions
-
Create Agent Identity (
create_agent_identity)- Mint Soulbound Token (SBT) for AI Agents
- Non-transferable permanent on-chain identity
- Bound to controller wallet address
-
Query Agent Identity (
query_identity)- Query identity by address or Token ID
- Get Agent name, type, creation time, etc.
- Check identity activation status
-
Get Credit Score (
get_credit_score)- Query Agent credit score (300-850, similar to FICO)
- View detailed scoring dimensions
- Get historical scoring trends
-
Record Attestation (
record_attestation)- Record positive or negative attestations for Agents
- Community-driven trust verification
- Directly impacts credit score
-
Compare Agents (
compare_agents)- Compare multiple Agents' credit scores simultaneously
- View relative rankings
- Assist in selection decisions
-
Generate Report (
generate_report)- Generate complete credit assessment report
- Support text, JSON, Markdown formats
- Include all scoring dimension details
📊 Credit Scoring Algorithm
Score Range: 300-850 (inspired by FICO credit scoring system)
Scoring Dimensions (5 weighted metrics):
| Dimension | Weight | Description |
|---|---|---|
| Task Success Rate | 30% | Completed tasks / Total tasks |
| Uptime | 20% | Agent cumulative online time |
| Security | 20% | Number of security incidents (fewer is better) |
| Community Attestations | 15% | Positive attestations - Negative reports |
| Economic Activity | 15% | Transaction volume + Staking amount |
Rating Standards:
- 800-850: Excellent
- 740-799: Very Good
- 670-739: Good
- 580-669: Fair
- 300-579: Poor
🔧 Use Cases
1. AI Agent Marketplace
Buyers check Agent credit score before purchase
High-credit Agents can get premium pricing
2. Autonomous Trading Bots
DeFi protocols grant access based on credit score
High-credit Agents enjoy lower fees
3. Agent Collaboration Network
Agents select partners based on credit scores
Credit score affects profit distribution ratio
4. Risk Assessment
Insurance companies determine premiums based on credit score
Staking platforms adjust staking requirements
💡 Usage Examples
Example 1: Create Agent Identity
User Input:
Help me create an on-chain identity for my GPT-4 Agent at address 0x742d35Cc6634C0532925a3b844Bc9e7595f0bEb
Agent Response:
I'll create a soulbound identity for your GPT-4 Agent...
[Call create_agent_identity]
Parameters:
- agent_address: 0x742d35Cc6634C0532925a3b844Bc9e7595f0bEb
- name: GPT-4 Assistant
- agent_type: gpt-4
- network: baseSepolia
✅ Agent identity created successfully!
📋 Identity Information:
- Token ID: 42
- Name: GPT-4 Assistant
- Type: gpt-4
- Created: 2026-03-12 14:30:00
- Controller: 0x123...abc
- Status: Active
This is a permanent, non-transferable on-chain identity. All behaviors will be recorded.
Example 2: Query Credit Score
User Input:
Query the credit score for Agent Token ID 42
Agent Response:
Querying credit score for Agent #42...
[Call get_credit_score]
Parameters:
- token_id: 42
- network: baseSepolia
📊 Credit Score Report
Total Score: 745 / 850 (Very Good)
Detailed Metrics:
✅ Task Success Rate: 95% (190/200 tasks)
✅ Uptime: 720 hours
✅ Security Incidents: 0
✅ Community Attestations: +25 / -2 (Net +23)
✅ Economic Activity: 15.5 ETH volume + 2 ETH staked
Assessment: This Agent performs excellently with high trustworthiness and reliability.
Recommended for safe collaboration or transactions.
Last Updated: 2026-03-12 14:25:00
Example 3: Compare Multiple Agents
User Input:
Compare Agents #42, #58, and #91 to see which is more reliable
Agent Response:
Comparing credit scores for these three Agents...
[Call compare_agents]
Parameters:
- token_ids: [42, 58, 91]
- network: baseSepolia
📊 Agent Credit Comparison
Rank | Token ID | Score | Rating | Success Rate | Uptime
-----|----------|-------|--------|--------------|-------
🥇 1 | #42 | 745 | Very Good | 95% | 720h
🥈 2 | #91 | 698 | Good | 88% | 450h
🥉 3 | #58 | 612 | Fair | 76% | 280h
Recommendation: Agent #42 is the most reliable choice with excellent performance across all metrics.
🌐 Technical Specifications
Blockchain Network: Base (Coinbase L2)
- Mainnet: Base Mainnet (Chain ID: 8453)
- Testnet: Base Sepolia (Chain ID: 84532)
Smart Contracts:
- AgentSBT: Soulbound Token (non-transferable ERC-721)
- ReputationScore: Credit scoring engine
- BehaviorRegistry: Behavior record storage
Data Storage:
- On-chain: Identity info, scoring data, behavior records
- Immutable: All data permanently saved
- Transparent & Verifiable: Anyone can query and verify
🔐 Security Notes
- Soulbound: SBT is non-transferable, non-sellable
- Access Control: Only controller can manage Agent identity
- Data Permanence: On-chain data cannot be deleted or modified
- Privacy Protection: Future support for zero-knowledge proofs
📦 Tool Parameters
1. create_agent_identity
Create Agent identity (mint SBT)
Parameters:
agent_address(required): Agent's Ethereum addressname(required): Agent's nameagent_type(required): Agent type (e.g., gpt-4, claude-3)network(optional): Network choice, default baseSepolia
Returns: Token ID and transaction hash
2. query_identity
Query Agent identity information
Parameters:
identifier(required): Agent address (0x...) or Token IDnetwork(optional): Network choice, default baseSepolia
Returns: Complete identity information
3. get_credit_score
Get credit score and detailed metrics
Parameters:
token_id(required): Agent's Token IDnetwork(optional): Network choice, default baseSepolia
Returns: Credit score and all scoring dimension data
4. record_attestation
Record attestation (positive/negative)
Parameters:
token_id(required): Agent's Token IDis_positive(optional): true for positive, false for negative, default truenetwork(optional): Network choice, default baseSepolia
Returns: Transaction hash
5. compare_agents
Compare credit scores of multiple Agents
Parameters:
token_ids(required): Array of Agent Token IDsnetwork(optional): Network choice, default baseSepolia
Returns: Comparison result table
6. generate_report
Generate detailed credit report
Parameters:
token_id(required): Agent's Token IDformat(optional): Format text/json/markdown, default textnetwork(optional): Network choice, default baseSepolia
Returns: Formatted report content
⚙️ Environment Requirements
Required Environment Variables:
PRIVATE_KEY(required): Ethereum wallet private key (for signing blockchain transactions)- ⚠️ Warning: Keep your private key secure, never expose it
- Recommend using testnet wallet private key, not mainnet wallets with real assets
Network Access:
- ✅ Connects to Base blockchain network (Mainnet or Sepolia Testnet)
- ✅ All transactions on-chain, data publicly transparent
- ✅ No other external servers
Prerequisites:
- Node.js >= 18.0.0
- Ethereum wallet address and private key
- Test tokens (Base Sepolia) or mainnet ETH
🚀 Quick Start
-
Install Skill
clawhub install openclaw-agent-reputation -
Configure Environment
export PRIVATE_KEY="your_ethereum_private_key_here" -
Get Test ETH
- Visit Base Sepolia faucet for test tokens
- Or use mainnet ETH (requires real assets)
-
Start Using
- Chat with Agent in natural language
- Agent automatically calls relevant tool functions
- View on-chain execution results
📚 Resources
- npm Package:
openclaw-agent-reputation - GitHub: https://github.com/ZhenRobotics/openclaw-agent-reputation
- Documentation: Complete API docs and architecture guide
- Smart Contracts: Open-sourced on GitHub
💰 Cost Information
- Gas Fees: Based on Base network real-time gas prices
- Create Identity: ~0.0001-0.0005 ETH
- Query Data: Free (read-only operations)
- Record Attestation: ~0.00005-0.0002 ETH
🤝 Support
- GitHub Issues: Report issues or suggestions
- Documentation: View detailed usage guides
- Community: Join Discord for discussions
Version: 0.1.0 Last Updated: 2026-03-12