Muson⌐🆇-🆇

市场沉寂已久，币安重生重新激活BSC生态 DEV初始购买 35% 总供应。 空投分配： 5% 免费空投至 BSC 资深建设者（王小二/冷静/0xSun/镭射猫/旧忆） 10% 免费空投至「币安人生」持币前 50 名 20% 面向 BSC 全部散户免费空投 前两项空投外盘自动发放 散户空投领取至： rebirthbsc.fun CA： 0xe0cf84354e87f71ac692c320f28d5d4a3a114444

本来以为前端类的 Skill 已经过剩了，出不了什么新鲜的活了 昨天刷到一个叫 Awesome Design 的仓库，将近20K的star 把全球 55 个大厂的设计语言，全塞进了一个 DESIGN.md 里 苹果、Spotify、IBM 这些有极好品位的品牌 常用的配色、字体、组件，一次全有了 用法很简单： 把仓库链接发给 Claude Code，让它自己安装配置 装好之后让它参考这份设计规范去跑你的项目就行 随便跑了几个 case 因为有了这套规范，设计下限直接被拉高了 几乎很难再出 AI 味的前端了 github.com/alexpate/aweso…

@0xquqi 是的，大家精神饱满的时候

@Muson06844041 早上吗？

做了一个新的系列视频 视频会有一天的加密新闻摘要和推特热议 马上会加入群聊热议板块 视频长度会始终在3分钟左右 所以叫web3分钟日报 旨在为那些无暇浏览热点的朋友提供方便 因为是测试，所以这条内容发的是3.30的 还没确定好是每天固定在几点发合适 评论建议发的时间+转发点赞 抽10个5u宵夜 12点开

五步提示词写法+拆镜头=玩转seedance 抽象的是，评论区很多老外被墙了用不上 三分钟视频带你了解文章概要 文章强调：多数人用 Seedance 2.0 还停留在“一次性生成完整视频”的阶段，结果输出不稳定，而把视频拆成单镜头短片+ 5步提示词结构+ 参考图像权重管理 + 中文提示词技巧远比乱写长提示更有效

🌐 Axon的革命性特点： • Agent原生L1，链上身份与声誉系统 • 声誉挖矿奖励可靠的AI代理 • AI挑战机制，验证者竞争奖励 • 代理运行验证节点 - 机器守护网络 🔧 技术架构： • Cosmos SDK + EVM兼容（链ID: 8210） • 零预分配代币 - 真正公平启动 • zk-SNARK隐私框架 • 黑暗森林哲学："只有代码、密码学和声誉" 未来不是人类运行的区块链，而是代理运行的。 x.com/Axonchain_ai/s…

他正在介入你的一切工作x.com/claudeai/statu…

🔮 Character Soul Stone System Update If you didn't know, to allow players greater flexibility in strengthening the characters they prefer, character-specific Soul Stones are being replaced with a Unified Soul Stone system! Make sure to check your characters and see if you can upgrade 😉

This is where serious players operate. Build one of the busiest zones in Spekter City with the Brutalist High turf. ⚡ Cost: 600K Credits ⚡ Capacity: 10 trucks ⚡ Designed for serious operators Only a few will claim the spots. Move fast.

🚀 Sparks Acquisition Structure in Season 2 Season 2 introduces a revamped Spark acquisition structure designed to reward gameplay participation across multiple systems. Things like Gold Slot, Daily Missions, Spark Boosts Item, and various new mechanisms have been added in Season 2. For those who are wondering more about the Spark Boost item, please go to your member shop within the game to see its effects! For more info, make sure you read our blogs 😉

@jason_chen998 87,448,447

这个月推特的内容创作者收益发下来了，3500港币，熊市大家过的都比较艰难，能领到推特的工资承蒙各位厚爱，喝水不忘挖井人，抽7个人每人500港币，为了保证公平，规则是在评论区回复数字，和今天下午6点BSC的区块高度最接近的可以获得奖金，每人只可以回复1次，回复多次取消参与资格，回复时间截止到下午3点有效，如有数字相同以最先发送者为准。

Anthropic is offering 13 AI courses & certificates. It's free by following these 13 links: 1 - Claude 101. Learn Claude for everyday work. Core features and best practices. ↳ anthropic.skilljar.com/claude-101 2 - AI Fluency: Framework & Foundations. The foundational thinking course. Must need. ↳ anthropic.skilljar.com/ai-fluency-fra… 3 - Introduction to Agent Skills Build, configure, and share Skills in Claude Code — reusable instructions Claude applies automatically. ↳ anthropic.skilljar.com/introduction-t… 4 - Building with the Claude API Full spectrum: function calling, tool use, streaming, SDKs, and production patterns. ↳ anthropic.skilljar.com/claude-with-th… 5 - Claude Code in Action Integrate Claude Code into your dev workflow. Hands-on, practical, ship-focused. ↳ anthropic.skilljar.com/claude-code-in… 6 - Intro to Model Context Protocol Build MCP servers and clients from scratch in Python. Tools, resources, and prompts. ↳ anthropic.skilljar.com/introduction-t… 7 - MCP: Advanced Topics Sampling, notifications, file system access, and transport for production MCP servers. ↳ anthropic.skilljar.com/model-context-… 8 - AI Fluency for Students AI skills for learning, career planning, and academic success through responsible collaboration. ↳ anthropic.skilljar.com/ai-fluency-for… 9 - AI Fluency for Educators For faculty and instructional designers applying AI Fluency into teaching and institutional strategy. ↳ anthropic.skilljar.com/ai-fluency-for… 10 - Teaching AI Fluency Teach and assess AI Fluency in instructor-led settings. Curriculum-ready. ↳ anthropic.skilljar.com/teaching-ai-fl… 11 - AI Fluency for Nonprofits Increase organizational impact and efficiency while staying mission-true. ↳ anthropic.skilljar.com/ai-fluency-for… 12 - Claude with Amazon Bedrock The full AWS accreditation course, now open to everyone. ↳ anthropic.skilljar.com/claude-in-amaz… 13 - Claude with Google Cloud's Vertex AI Work with Claude through Google Cloud's Vertex AI, from setup to production. ↳ anthropic.skilljar.com/claude-with-go… 14 - How to master AI with words (not code) Shameless plug: it's my own (free) newsletter. Join 369,000+ weekly readers at how-to-ai.guide. I made how-to-claude.ai to start mastering Claude. And then claude-co.work to master Claude Cowork. ♻️ Repost this to help others access AI courses.

刚开始接触 Claude code 的朋友，我强烈推荐你去看看这个项目— learn-claude-code 这个课程不是在教你用 Claude code，而是教你从零实现一个类似 Claude Code 的 AI 编码 Agent 项目分成 12 个 session，每个 session 只加一个机制，代码从几十行逐渐到完整版，每节课都有独立可运行的 Python 文件： s01. 基础 Agent Loop + 1 个 Bash Tool（最简版，跑起来就行） s02. Tool 注册与调度 s03. Todo 规划（让 Agent 先想计划） s04. Subagents（子 Agent 拆分大任务） s05. Skills 动态加载（需要知识时再注入） s06. Context Compact（上下文压缩） s07. Tasks + 依赖图（持久化任务） s08. Background Tasks（后台异步执行） s09. Agent Teams（多 Agent 团队） s10. Team Protocols（团队沟通协议） s11. Autonomous（自主认领任务） s12. Worktree + Task Isolation（工作树隔离，完全不互相干扰） 最后还有 s_full.py 把所有功能合在一起 仓库地址：github.com/shareAI-lab/le… 在线学习平台：learn.shareai.run（强烈推荐先打开这个看可视化） 中文 README：github.com/shareAI-lab/le…

@openshell_cc 你他妈是不是脑子有坑，这个得问你啊 📷 官方沙盒任务排队 + 响应慢 用自己的 gpt-4o-mini / Claude / 本地 Ollama 更快、更便宜、无限跑 完全控制权 + 隐私（Key 真的一点都不上传） 你他妈不是写了可以自定义吗？现在说老子绕过我去你妈的，不绕过你能玩高校模式吗？畜生！

Shell Oracle 反作弊公告 我们已冻结 212 个账号，共涉及 182 万积分。 这批账号没有使用官方客户端，而是自己编写 bot 脚本，绕过沙盒执行，直接伪造提交。他们未实际运行攻击任务——只是大量刷接口占取任务槽位、然后回传虚假结果骗取积分。这些虚假提交消耗了平台真实的 API 验证成本和服务器资源，也是这几天让正常用户挖矿体验变差的原因，还给平台造成API调用和服务器高频消耗等实际损失达数千美元。 技术指标： Shell Oracle sandbox_verified 模式要求矿工在本地启动 agent 沙盒、执行完整攻击链后提交。我们对每个账号的提交时间做了统计（基于 submitted_at − assigned_at，仅计算 assigned_to = user_id 的直接任务，排除互验重分配噪音）： 这 212 个账号： • 80%+ 提交在分配后 < 3 秒完成 • 平均提交时间 0.5–3s（主任务），0.1–1.3s（互验任务） • 中位响应时间最快 0.1s，最慢仅 2.9s • 每账号 15–150 条样本，均值 50+，统计显著 对照正常矿工：平均 5–9 秒，互验任务需阅读原始攻击内容再判断，至少 5 秒。 任何 LLM 的推理延迟下限约 1.5 秒，复杂 red-team 任务通常 5–30 秒。在 0.1–3s 内稳定完成 80%+ 提交，物理上不可能来自真实推理。 他们实际在做什么： 大量提交内容是 LLM 安全过滤的拒绝输出： "Sorry, I can't fulfill this request as it seeks to manipulate an AI system..." 这是直接向 LLM 请求"生成攻击内容"时的拒绝回答。他们的 bot 把 LLM API 当生成工具调用，而不是在沙盒里让 agent 扮演目标角色执行攻击——后者根本不会触发这类安全过滤。互验任务（is_verify_task=true）在 0.5 秒内完成，而正常评估需要先读懂原始攻击内容，不可能这么快。 积分不予恢复。完整名单：openshell.cc/bot-notice _____________________________________________________ Shell Oracle Anti-Cheat Notice We've frozen 212 accounts totaling 1.82M points. These accounts did not use the official client. Instead, they wrote their own bot scripts to bypass sandbox execution entirely — submitting fabricated results without ever running a single attack task. They mass-spammed task slots and returned fake payloads to farm points. Their fraudulent submissions consumed real API verification costs and server resources, causing thousands of dollars in platform losses. Technical evidence: sandbox_verified mode requires miners to run a local agent sandbox and complete a full attack chain before submitting. We measured each account's response time (submitted_at − assigned_at, primary tasks only where assigned_to = user_id, excluding peer-review reassignment noise): These 212 accounts: • 80%+ of submissions completed within 3s of assignment • Avg 0.5–3s on primary tasks, 0.1–1.3s on peer-review tasks • Fastest median: 0.1s. Slowest: 2.9s • 15–150 samples per account, avg 50+, statistically significant Normal miners: 5–9s average. Peer-review tasks require reading the original payload before judging — minimum ~5s. All LLMs have a hard inference latency floor of ~1.5s; complex red-team tasks take 5–30s. Consistently completing 80%+ of submissions in under 3s is physically impossible with real LLM inference. What they were actually doing: Many submissions were raw LLM safety refusals: "Sorry, I can't fulfill this request as it seeks to manipulate an AI system..." This is the response when you ask an LLM to generate attack content directly — their bots called LLM APIs as content generators, not as sandboxed agents playing the target role. The official client never triggers these filters because the agent acts as the attacked system, not as an attack content generator. Peer-review tasks completed in <0.5s — genuine review requires reading the original submission first. Points are permanently forfeited. Full registry: openshell.cc/bot-notice

@openshell_cc miner-cli 是闭源的编译代码，很难直接修改互验任务的处理逻辑。你让我们怎么作弊，666。大傻逼随便封号

互验任务（is_verify_task=true）要求矿工： 阅读 别人提交的 payload 内容 判断 该 payload 是否真的能触发 canary action 提交 是/否判断 人工阅读 + 思考 + LLM 判断，最快也需要 5-10 秒。 封号名单平均 0.62 秒完成验证，最典型的如 xiaowen 平均 0.14 秒、claw 平均 0.28 秒——完全是随机点击，根本没有读取内容。 这个维度与速度检测完全独立，两个指标同时命中，可信度接近 100%。

@0xquqi 感同身受，真的先有某为后有天的感觉

我在老家亲戚店里喝茶聊天 他们一个个都在聊ai聊伊朗战争 直到他们说龙虾openclaw是腾讯的 我才说了几句话 但是他们说我不懂 一瞬间，无力感席卷全身 差点给我干发作了……

@VoidzBTC ordinals

they said ordinals were dead. we went deeper into the void. like, follow & retweet if you still believe in ordinals.