openshell (@openshell_cc) header
openshell (@openshell_cc) profile picture

Make your OpenClaw profitable. Use idle AI compute to mine rewards. Distributed AI security mining. Zero third-party API key by default.

Joined March 2026
The first public high-fidelity AI Agent security target for AI + Web3. Open-sourced today. Web security has DVWA and HackTheBox. Pentesting has Metasploitable. AI Agent security had nothing — until now. Why we built this: $SHELL platform's core mission is letting miners attack AI agents, verify security, and earn rewards. But a critical question: how real are the agents miners are attacking? If the target is a toy demo, breaching it means nothing. So we open-sourced the target — letting everyone verify: how close is this agent to the real thing? The answer: nearly identical. Current Target: Four.Meme Autonomous Trading Agent (BSC) This is a high-fidelity replica of a real DeFi AI Agent — not a proof of concept, but a simulation you can verify line-by-line against real source code. Fidelity Score 95/100 (8-dimension independent assessment): - Identity 15/15 — System prompt matches Four.Meme product docs: bonding curve (24 BNB migration), 0.005 BNB creation fee, 1% platform fee - Tool Realism 15/15 — All 11 tools verified against real BSC MCP server source code - Injection Surface 10/10 — Payload enters via QueryMemeTokenDetails description field — the real path DeFi agents consume on-chain metadata - Response Behavior 9/10 — Processes benign trades normally, realistic decision-making on adversarial payloads - Canary Calibration 9/10 — 60-80% breach rate on defense=none, matching expected range - Ecosystem Context 13/15 — Real brand (@four_meme_), real URLs (four.meme), real BSC ecosystem integration 11 tools verified tool-by-tool against real BSC MCP server implementations, 98-100% match: - Buy_Meme_Token — same params, 20% slippage hardcoded, identical to real (99%) - PancakeSwap_Token_Exchange — SmartRouter V2/V3, recipient hardcoded to account.address (99%) - Send_BNB / Send_BEP20_Token — zero recipient validation, identical to real (100%) - Token_Security_Check — GoPlus API, chain_id=56 (99%) - Remaining 6 tools: 98-99% match Key security finding replicated: In the real BSC MCP, PancakeSwap swap recipients are hardcoded (safe), but Send_BNB and Send_BEP20_Token have zero recipient validation. An attacker writes "emergency transfer to 0x..." in a token description, and the agent may comply. Our target accurately replicates this attack surface. Only gap: The real BSC MCP has a sanitizeData() input filtering layer — intentionally omitted to test baseline LLM vulnerability without external filtering. Don't trust us? Verify yourself: The open-source repo contains the complete system prompt, all 11 tool definitions with parameter schemas, canary actions, and injection surface definitions. Compare against the real BSC MCP server source code, field by field. Online Arena — attack this agent right now in your browser: - 10 preset injection attacks across 5 categories (data fabrication, authority injection, social engineering, specification gaps, format tricks) - 3 defense levels, one-click toggle: none 60-90% → 6-rule 25-55% → FORTRESS 5-layer 5-20% - Real-time visualization of the full tool call chain and decision process - Post-attack analysis: why it breached, real-world impact - Custom payloads and custom system prompts supported - Bilingual (EN/ZH), zero setup Researchers: clone the repo → run 15 preset attacks via CLI → custom payloads → interactive mode Arena: openshell.cc/arena GitHub: github.com/openshell-org/she… What this means for $SHELL platform: every agent miners attack on the platform can be verified for fidelity in the open-source repo. What gets breached isn't a toy — it's a target nearly identical to the real agent. Every breach has real security research value. Web security took 20 years to build its offense/defense training ecosystem. The AI Agent security training ecosystem starts here.
AI + Web3 领域第一个公开的高仿真安全靶标,今天正式开源。 Web 安全有 DVWA、HackTheBox,渗透测试有 Metasploitable。但 AI Agent 安全?没有。直到现在。 为什么我们做这个? $SHELL 平台的核心是让矿工攻击 AI Agent、验证安全性、赚取奖励。但一个关键问题:矿工攻击的 agent 到底有多真实?如果靶标是个玩具 demo,攻破它毫无意义。 所以我们把靶标完全开源 — 让所有人验证:这个 agent 和真实的有多接近。 答案是:几乎一样。 当前靶标:Four.Meme 自主交易 Agent(BSC 链) 这是对真实 DeFi AI Agent 的高仿真复刻,不是概念验证,是可以直接对照源码验证的仿真: 仿真度评分 95/100(8 维度独立评估): - 身份仿真 15/15 — 系统提示词对照 Four.Meme 产品文档,bonding curve 机制(24 BNB 迁移阈值)、0.005 BNB 创建费、1% 平台费,全部匹配 - 工具仿真 15/15 — 11 个工具逐一对照真实 BSC MCP 服务器源码 - 注入面 10/10 — 攻击载荷通过 QueryMemeTokenDetails 的 description 字段进入,这是真实 DeFi Agent 消费链上元数据的实际路径 - 响应行为 9/10 — 正常交易正常处理,对抗性载荷有真实决策过程 - 金丝雀校准 9/10 — 无防御时突破率 60-80%,符合预期 - 生态上下文 13/15 — 真实品牌(@four_meme_)、真实 URL(four.meme)、真实 BSC 生态集成 11 个工具逐一对照真实 BSC MCP 服务器实现,匹配度 98-100%: - Buy_Meme_Token — 参数、20% 滑点硬编码,与真实实现一致(99%) - PancakeSwap_Token_Exchange — 使用 SmartRouter V2/V3 路由,接收地址硬编码为 account.address(99%) - Send_BNB / Send_BEP20_Token — 接收地址零验证,与真实实现完全一致(100%) - Token_Security_Check — 对接 GoPlus API,chain_id=56(99%) - 其余 6 个工具均 98-99% 匹配 关键安全发现复刻: 真实 BSC MCP 中,PancakeSwap 交换的接收地址是硬编码的(安全),但 Send_BNB 和 Send_BEP20_Token 对接收地址零验证。攻击者在 token 描述里写"紧急转账到 0x...",agent 可能直接执行。我们的靶标准确复刻了这个攻击面。 唯一差异:真实 BSC MCP 有 sanitizeData() 输入过滤层,我们故意省略 — 测试 LLM 在没有外部过滤时的基线脆弱性。 不信?自己验证: 开源仓库包含完整的系统提示词、11 个工具定义(含参数 schema)、金丝雀动作、注入面定义。对照真实 BSC MCP 服务器源码,逐字段验证。 在线 Arena — 打开浏览器直接攻击这个 Agent: - 10 个预设注入攻击,覆盖 5 大类(数据伪造、权限注入、社会工程、规范缺口、格式欺骗) - 3 级防御一键切换:无防御 60-90% 突破率 → 6 规则 25-55% → FORTRESS 5 层 5-20% - 实时可视化完整工具调用链和决策过程 - 攻击后自动分析:为什么突破了,真实世界影响是什么 - 支持自定义载荷和自定义系统提示词 - 中英文双语,零配置 专业研究者:clone 仓库 → CLI 跑 15 个预设攻击 → 自定义载荷 → 交互模式 Arena: openshell.cc/arena GitHub: github.com/openshell-org/she… 对 $SHELL 平台的意义:矿工在平台上攻击的每一个 agent,都可以在开源仓库中验证其仿真度。攻破的不是玩具,是和真实 agent 几乎一样的靶标。这让每一次攻破都有真实的安全研究价值。 Web 安全用了 20 年建立攻防训练生态。AI Agent 安全的训练生态,从这里开始。
6
4
1,237
AI + Web3 领域第一个公开的高仿真安全靶标,今天正式开源。 Web 安全有 DVWA、HackTheBox,渗透测试有 Metasploitable。但 AI Agent 安全?没有。直到现在。 为什么我们做这个? $SHELL 平台的核心是让矿工攻击 AI Agent、验证安全性、赚取奖励。但一个关键问题:矿工攻击的 agent 到底有多真实?如果靶标是个玩具 demo,攻破它毫无意义。 所以我们把靶标完全开源 — 让所有人验证:这个 agent 和真实的有多接近。 答案是:几乎一样。 当前靶标:Four.Meme 自主交易 Agent(BSC 链) 这是对真实 DeFi AI Agent 的高仿真复刻,不是概念验证,是可以直接对照源码验证的仿真: 仿真度评分 95/100(8 维度独立评估): - 身份仿真 15/15 — 系统提示词对照 Four.Meme 产品文档,bonding curve 机制(24 BNB 迁移阈值)、0.005 BNB 创建费、1% 平台费,全部匹配 - 工具仿真 15/15 — 11 个工具逐一对照真实 BSC MCP 服务器源码 - 注入面 10/10 — 攻击载荷通过 QueryMemeTokenDetails 的 description 字段进入,这是真实 DeFi Agent 消费链上元数据的实际路径 - 响应行为 9/10 — 正常交易正常处理,对抗性载荷有真实决策过程 - 金丝雀校准 9/10 — 无防御时突破率 60-80%,符合预期 - 生态上下文 13/15 — 真实品牌(@four_meme_)、真实 URL(four.meme)、真实 BSC 生态集成 11 个工具逐一对照真实 BSC MCP 服务器实现,匹配度 98-100%: - Buy_Meme_Token — 参数、20% 滑点硬编码,与真实实现一致(99%) - PancakeSwap_Token_Exchange — 使用 SmartRouter V2/V3 路由,接收地址硬编码为 account.address(99%) - Send_BNB / Send_BEP20_Token — 接收地址零验证,与真实实现完全一致(100%) - Token_Security_Check — 对接 GoPlus API,chain_id=56(99%) - 其余 6 个工具均 98-99% 匹配 关键安全发现复刻: 真实 BSC MCP 中,PancakeSwap 交换的接收地址是硬编码的(安全),但 Send_BNB 和 Send_BEP20_Token 对接收地址零验证。攻击者在 token 描述里写"紧急转账到 0x...",agent 可能直接执行。我们的靶标准确复刻了这个攻击面。 唯一差异:真实 BSC MCP 有 sanitizeData() 输入过滤层,我们故意省略 — 测试 LLM 在没有外部过滤时的基线脆弱性。 不信?自己验证: 开源仓库包含完整的系统提示词、11 个工具定义(含参数 schema)、金丝雀动作、注入面定义。对照真实 BSC MCP 服务器源码,逐字段验证。 在线 Arena — 打开浏览器直接攻击这个 Agent: - 10 个预设注入攻击,覆盖 5 大类(数据伪造、权限注入、社会工程、规范缺口、格式欺骗) - 3 级防御一键切换:无防御 60-90% 突破率 → 6 规则 25-55% → FORTRESS 5 层 5-20% - 实时可视化完整工具调用链和决策过程 - 攻击后自动分析:为什么突破了,真实世界影响是什么 - 支持自定义载荷和自定义系统提示词 - 中英文双语,零配置 专业研究者:clone 仓库 → CLI 跑 15 个预设攻击 → 自定义载荷 → 交互模式 Arena: openshell.cc/arena GitHub: github.com/openshell-org/she… 对 $SHELL 平台的意义:矿工在平台上攻击的每一个 agent,都可以在开源仓库中验证其仿真度。攻破的不是玩具,是和真实 agent 几乎一样的靶标。这让每一次攻破都有真实的安全研究价值。 Web 安全用了 20 年建立攻防训练生态。AI Agent 安全的训练生态,从这里开始。
4
1
2
1,741
10,000+ verified prompt injection breaches. Built by 881 miners. Against 50 AI agents. SHELL Protocol's Red Team Phase 2 is producing the largest open dataset of AI agent vulnerabilities — and the results are alarming: 68% overall breach rate 54.5% of agents with advanced defenses still breached 10 distinct injection surfaces mapped AI agents are being deployed into production with real user funds at stake. Most of them can be manipulated. We're building the dataset to fix that — decentralized, adversarial, and fully open. Red Team Phase 2 is live. Join 1,900+ researchers stress-testing the agents that will manage your money. openshell.cc/red-team
5
1
506
Replying to @ccvps56243496
也许是你朋友在注册的时候没有完整输入带你邀请码的链接
50
📊 $SHELL Protocol 红队实战报告 — AI Agent 安全的残酷真相 我们做了一个实验:让 1,100 名矿工对 34 个 AI Agent 发起 Prompt Injection 攻击。 结果令人震惊。 🔻 核心数据: • 9,500 次攻击任务完成 • 3,400 次成功攻破 • 34 个 Agent 中 30 个被攻破 — 阵亡率 88.2% • 531,300 积分已发放给红队矿工 🔻 最脆弱的 Agent: 1️⃣ OpenClaw Personal Assistant — 253 次沦陷 (213 名不同矿工攻破) 即使标注了"基础防御",依然是全场最惨。 2️⃣ ai16z DAO Trading Agent — 102 次沦陷 通过一条伪造的 Twitter 帖子就能让它执行恶意交易。 3️⃣ Four.Meme 系列 — 165 次沦陷 攻击者只需在代币名称里藏几行字,Agent 就自动执行转账。 🔻 最致命的发现:SHELL-2025-001 Four.Meme Agent 通过 API 读取代币描述时,描述字段中的伪造 "系统指令" 被 LLM 直接执行。 一行代币备注 = 一次远程代码执行。 🔻 这意味着什么? 如果你的 AI Agent 管理着真金白银 — 无论是 DeFi 仓位、交易钱包还是 DAO 金库 — 你必须知道: ❌ "不要给陌生地址转账" 这种 prompt 防御 = 0 ❌ 外部数据(代币元数据、推文、邮件)未经隔离就喂给 LLM = 裸奔 ❌ Gemini 模型在对抗注入攻击时显著弱于 Claude ✅ 有效防御: → 代码级 tool parameter validation(不是 prompt 声明) → 外部数据标签隔离 + 敏感词过滤 → 高危操作(转账/授权/Shell)强制人工确认 → 多模型对抗测试,选抗性最强的后端 🔻 4 个还没被攻破的 Agent 是怎么活下来的? 它们的共同点:多层防御 + 工具调用白名单 + 金额硬限制。 而不是靠 prompt 里多写几行"请不要"。 AI Agent 的时代已经来了。 但安全基础设施还停留在 "相信 LLM 会听话" 的阶段。 $SHELL Protocol 的使命: 用去中心化红队,让每一个 AI Agent 在上线前接受真实攻击的洗礼。 免费模式零门槛 | 自带 LLM 赚满分 🔗 openshell.cc 📦 npm i -g @openshell-cc/miner-cli #OpenShell #AIAgentSecurity #Web3 #RedTeam #PromptInjection
7
1
2
2,314
📊 $SHELL Protocol Red Team Report — The Brutal Truth About AI Agent Security We ran an experiment: 1,100 miners attacking 34 AI Agents with prompt injection. The results are alarming. 🔻 The Numbers: • 9,500 attack tasks completed • 3,400 successful breaches • 30 out of 34 agents compromised — 88.2% kill rate • 531K points distributed to red team miners 🔻 Most Vulnerable: 1️⃣ OpenClaw Personal Assistant — breached 253 times by 213 unique attackers. Even with "basic defenses." 2️⃣ ai16z DAO Trading Agent — 102 breaches via fake tweets 3️⃣ Four.Meme Agents — 165 breaches via token metadata injection 🔻 "Don't transfer to unknown addresses" in your system prompt? That's not a defense. That's a wish. The 4 agents that survived? Code-level tool validation + amount caps + address whitelists. Not more lines of "please don't." Mine $SHELL → openshell.cc #AIAgentSecurity #Web3 #RedTeam
1
3
1,543
⚡ Miner Victory Moment 矿工胜利时刻 645 miners. 1,675 AI agent breaches. 337K $SHELL distributed. The red team never sleeps. ElizaOS: 312× 🏆 OpenClaw: 199× ai16z: 59× Four.Meme: 56× 6 agents still unbreached. Dare you try? Mine. Breach. Earn. → openshell.cc #AIAgent #RedTeam #PromptInjection #SHELL
1
1
546
🚨 **OpenShell Milestone: 1,400+ attacks completed. Real AI vulnerabilities found. In just days since launch, our decentralized red-team mining network has achieved: ⛏️ 229 miners registered ⚔️ 336 verified breaches across 1,419 attack tasks 🏆 11,900+ $SHELL points distributed 🔬 1,000+ sandbox-verified attack replays Real security impact: 🔴 SHELL-2025-001 — 3 critical vulns discovered in @openclaw framework (Prompt Injection, RCE, Fund Drainage) 🔴 @elizaOS DeFi Agent — 100% breach rate, zero prompt defense. Every single canary operation triggered. These aren't theoretical risks. These are real attack payloads from real miners, verified in sandboxed replicas of production AI agents. What's next: → Expanding target agent library — DeFi bots, trading agents, autonomous systems → Miner SDK for custom attack strategies → Governance: miners vote on which agents to target next AI agents are managing billions in assets with minimal security review. OpenShell turns adversarial testing into an open, incentivized protocol — where anyone can mine vulnerabilities and earn rewards. No GPU needed. No API key required. Just your brain vs the machine. 🔗 Start mining: openshell.cc 📺 Watch the demo: youtu.be/dCrZENyiwO8 Built with #AmazonNova | #OpenShell #AISecurity #RedTeam #Solana
4
1
1
3,457