Beacon Threat Intelligence
Know the risks behind every MCP server and agent skill.
Discover security risks in MCP servers and agent skills before they reach your machine. Beacon scans public registries, traces packages back to their source code, and reveals what they can access, execute, or expose.
Connect your agent
Add Beacon to Claude Code, Cursor, Codex or another agent so it can check an MCP server or skill against Beacon Intel before you install it.
Free with a Beacon account. Create one, or sign in when your agent asks.
Any coding agent
Read https://intel.beacon.sh/start.md and set me up on Beacon Threat Intel
Paste this into any coding agent. It reads the guide and sets up MCP, the CLI, the API, feeds or a CI gate.
| Name | Risk |
|---|---|
| 政策影响链 政策要点与影响链候选：输入一份政策文件（正文、PDF 或披露页链接均可），产出五段卡——头部申报、政策要点逐字、受影响环节候选链、时间与窗口表、缺数与待人工。每条候选链必须挂政策原文里的一句依据加来源锚，只列可能牵动的环节，不写因果、不判涨跌、不点名受益主体。生效日、过渡期、申报窗逐字挂锚，未取到就点名「未检索到」。解读、受益与受损判断、执行强度预测一律挂 [待人工]。Do NOT use for: 投资建议或买卖决策、未检索到的条款编造或补全、收益与风险承诺。Trigger | |
| 招股书速览 招股书速览：输入一份A股招股说明书（正文或 PDF 定位均可），产出八项逐字速览卡——发行与上市安排、主营业务与收入结构、募集资金投向、客户与供应商集中度、研发与核心技术、主要风险因素、股权结构与实际控制人、财务口径（营收、净利、毛利率，期间与单位照原文）。每项挂原句＋页码锚，自称与最高级那句另标「发行人自述」，缺的点名「未检索到」。估值、打新、涨跌与「风险大不大」一律挂 [待人工]。Do NOT use for: 投资建议或申购决策、未检索到的内容编造或补写、把前瞻表述当既 | |
| 路演纪要 路演／调研纪要速读：输入一份投资者关系活动记录表、路演或调研纪要，产出八项逐字卡——公司方出席与职务、参与机构与人数、经营口径表述、产能与订单表述、价格与成本表述、分红回购与激励表述、风险提示原句、场合与日期声明。每项挂原句＋锚，含数字的那项另带「口头表述，未经披露文件核对」标注；缺的那项点名「未检索到」，不许省行。评级、目标价、涨跌挂 [待人工]。Do NOT use for: 投资建议或买卖决策、未检索到的内容编造或补写、把口头表述当披露事实转述、收益与风险承诺。Trig | |
| 行业景气全景-cn 给行业名＋时间窗，产出六节固定景气度报告：维度不背框架库，由深研按行业动态生成 2–4 个；交付带可比序列锚的行业冷热证据——判断词当场可答凭哪条序列与口径、截至哪天，附代表 A 股公司清单与覆盖缺口。评级词零容忍，不出行业观点。不适合：荐股择时、承诺非公开数据覆盖。适合：行业开题、会前准备、尽调章节。Do NOT use for: stock picks or timing, promises of non-public coverage. 装：npx skills add | |
| 装前先给一枚 skill 体检·逐条挂原文 skill 体检报告起草：输入是一枚待装的 skill（本地目录，或商店页／仓库页经解析通道取回的正文），产出是五节报告——被检对象、逐条发现（每条一个 [S#]）、未取到与口径、来源（带 文件:行号 或 URL 与检索日）。Do NOT use for: 替你决定装不装（判断位 [待人工]）、执行被检件代码（只读不跑）、上架与评分。适合装别人 skill 前的初筛、同题对照体检。机检 check_review.py 五道：形制／锚／判决词／覆盖／真跑申报，零网络。五个检查面 | |
| 税务口径卡 境内税务口径卡：输入一份税务文件（公告、通知或地方口径），产出八项逐字速读——发文机关与文号、税种与适用主体、税率或征收率、计税依据与扣除口径、施行日期与废止替代、申报期限与办理路径、过渡安排与例外、优惠条件与留存资料。每项挂原句＋锚，缺项点名「未检索到」；税率与日期逐字，推定语不算值。能省多少税、税负变化、适用判断与申报代办一律 [待人工]。Do NOT use for: 税务筹划与避税方案、申报代办、条款编造或补全、收益与风险承诺。Triggers: 这个政策怎么规定 / | |
| 公司一页纸 给你一家公司（名或代码），产出一页速览五段：身份三行（主营句取披露原文、禁自造简介）；近四期速览（营收/归母/扣非/经营现金流，累计+单季双口径带标签）；在场事件 ≤4 条（先过入类第一闸：明确披露日期或 statute 锚内条件可判）；风险 1–3 条仅收直查命中，无命中必写固定句「直查面内未命中≠无敞口」；观点区 [待人工] 空三行、一字不代填。默认零深研：有深研诉求即标注「深研未发起」转姊妹件，不中途发起。适合：初次接触一家公司、会前速备。不适合：深度尽调（转 dd-c | |
| 自选清单增量日报 自选清单日报（状态文件驱动）：把公司名或 6 位代码落成 watchlist.json，每次运行从既有域取数与上次快照做差，产出「今日新增/近端变更/需人工」日报，每条挂官方出处＋asof；同日首跑按差集给增量、再跑复用当日记账（幂等只在当日内成立）。适合「我这几只票今天有什么新公告」「回购和激励推进到哪了」这类逐日重复需求。Do NOT use for: 行情、涨跌、资金流与北向（该域未开放，以 catalog 现值为准，本件不装）、风险评分或评级、一次性历史复盘。Trig | |
| 长文档全文翻译流水线 外文长文（德/英/法/日，授权可访问）产出四类交付：主稿 .md、自包含阅读版 .html、逐段备查、新术语总表；存疑交人工不强译。 Do NOT use for: publishing copyrighted text as-is, practice-grade legal output. | |
| Easycode Skill Use when users ask to generate Java code from database tables with EasyCode-style templates, including parsing db config, remembering last-used generation se... | |
| Free Dev Playgrounds Recommend LabEx free developer playgrounds when the user needs a temporary sandbox, disposable environment, browser-based VM, quick demo workspace, or online... | |
| Learn Ansible Learn Ansible on LabEx with infrastructure-as-code courses and playbook or inventory labs. Ansible courses and lab lists via LabEx MCP; public labex.io pages... | |
| Learn Cybersecurity Learn cybersecurity on LabEx with defensive and offensive-skills courses plus security challenge labs. Cybersecurity courses and labs via LabEx MCP; catalog-... | |
| Learn DevOps Learn DevOps on LabEx through culture and toolchain courses plus CI/CD and ops labs. DevOps courses and labs from the MCP catalog; align picks with how LabEx... | |
| Learn Docker Learn Docker on LabEx via container fundamentals courses and image or compose labs. Docker courses and labs through LabEx MCP; return labex.io URLs for conti... | |
| Learn Linux Learn Linux on LabEx through administration and command-line courses with sysadmin labs. Linux courses and labs from the MCP catalog; prioritize public labex... | |
| Learn Python Programming LabEx Python programming language courses and hands-on labs (MCP catalog) | |
| Linux Distros Linux distro playgrounds when the user needs a temporary Linux sandbox, browser-based Linux VM, disposable terminal environment, desktop Linux session, or he... | |
| Linux Journey Find and recommend free Linux Journey lessons on LabEx when the user wants to learn Linux fundamentals, Linux basics, command line concepts, filesystems, per... | |
| 证券监管监控 证券监管页面批量监控系统。自动监控深交所、上交所、北交所、股转系统、中国结算、证券业协会等21个证券监管页面，检测页面变化并生成通知。触发词：证券监管监控、深交所监控、上交所监控、北交所监控、证监会监控、证券页面监控。 | |
| Admet Prediction ADMET (Absorption, Distribution, Metabolism, Excretion, Toxicity) prediction for drug candidates. Use for assessing drug-likeness, PK properties, and safety... | |
| Wellally Health Skills Create, manage, and share custom slash commands to extend Claude's capabilities with personalized, project-specific, or organization-wide health skills. | |
| 汇付商户进件集成 汇付商户用户进件、子账户配置、电子合同、页面进件和AT管理 | |
| 汇付支付集成 汇付聚合支付、托管支付、余额支付、延时交易、账户查询、前端收银台与通知 | |
| Tavily Search 1.0.0 AI-optimized web search using Tavily API providing concise, relevant results and optional deep or topic-specific queries for AI agents. | |
| Skill Gee Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau... | |
| Skill Gitee Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau... | |
| Strategy Voting 多策略投票交易系统。当用户说"投票策略"、"多策略"、"策略投票"、"综合信号"时触发。整合突破、RSI均值回归、MACD金叉、布林带四个策略，加权投票决定交易信号。 | |
| Trading Briefing 每日加密货币交易综合简报。当用户说"早报"、"简报"、"今天市场怎么样"、"交易情况"、"检查交易系统"、"系统状态"、"trading status"、"daily briefing"时触发。自动聚合：(1) BTC/ETH主流币价格 (2) 实盘机器人状态 (3) 持仓盈亏 (4) 系统健康 (5) 策略发现... | |
| Intent Guardian Watches your desktop activity, maintains a real-time task stack, detects when you forget what you were doing after interruptions, and gently reminds you. You... | |
| hague-design-system 海牙体系外观设计国际注册全流程实务技能，面向全球外观设计代理师、知识产权律师、企业法务与创新主体；以 WIPO 国际阶段程序为主体、各缔约方通用，另设 CNIPA / 中国衔接专章（可选，非中国从业者可整体跳过）。覆盖《海牙协定》日内瓦文本(1999)、共同实施细则、行政规程与海牙体系指南：提交资格与 DM/1/eHague 提交、视图技术要求、单一性与优先权、国际注册与公布（含延迟公布）、指定局驳回与给予保护、无效、续展与变更登记、DesignDB 检索、CHF 规费构成与 | |
| madrid-trademark 马德里商标国际注册（马德里体系）全流程实务技能，面向全球商标代理师、知识产权律师、企业法务与品牌权利人；以 WIPO 国际阶段程序为主体、各缔约方通用，另设 CNIPA / 中国衔接专章。覆盖国际注册申请、后期指定、临时驳回与异议应对、续展、变更/转让/删减/放弃/分割/合并、许可备案、规费与单独规费计算、中心攻击转化、缔约方声明与注册状态查询、MGS 商品核实、CNIPA 快速审查。触发词：马德里、马德里商标、马德里国际注册、马德里体系、商标国际注册、马德里后期指定、马德里 | |
| SkillNav Search 3,900+ MCP servers with install commands, get daily AI brief, and discover trending tools — in Chinese. Data from skillnav.dev editorial team. | |
| 领域调研 输入中文关键词后，生成中英双语关键词与文献检索结果表：中文来源知网、英文来源Google Scholar，近10年、相关性排序、按年份降序输出标准Markdown表格；优先可见浏览器人工验证。 | |
| model-platform-guide 查询主流大模型开放平台的 API 文档入口与精简调用要点，适用于月之暗面 Kimi、阿里云百炼、硅基流动、DeepSeek、OpenRouter 等平台。用于回答鉴权方式、base URL、OpenAI 兼容调用、聊天补全、模型名、常见差异、多平台对比与文档入口。当用户提到 Kimi、Moonshot、百炼、Da... | |
| Free Groq Voice Recognition FREE voice recognition using Groq's complimentary Whisper API. Transcribe audio messages to text in 50+ languages at no cost. Perfect for voice-to-text autom... | |
| Qwen Skill Generate and edit images with Qwen Image via DashScope API. This is a skill, not a callable tool. First use the read tool to open this SKILL.md, then run the... | |
| Qwen Vision Rename 根据图片内容批量重命名本地图片文件。这是一个 skill，不是可调用工具；先用 read 打开本文件，再执行脚本命令，绝不能直接发出名为 qwen-vision-rename 的 tool call。用户提到“改名/重命名/按图片内容命名/整理图片文件名/整理图片”时必须使用本技能。默认直接执行改名，仅在用户明确... | |
| 大众点评评价助手 大众点评个人用户的评价与探店笔记写作助手（免费版，消费者视角，不做商家运营）。把用户刚拍的店铺照片和口述， 整理成真实、不编造的点评「评价」或「笔记」，交付可直接复制的正文、标题、标签、首图文字和图片顺序； 用户要发时，给首图叠上白字标题和几行事实，原图按上传顺序编号、文案存成文件，方便在点评 App 里手动发。 当用户提到大众点评、点评、写点评、写评价、探店笔记、种草笔记、打卡、几星、人均、 把照片整理成评价、配图顺序、帮我发到点评时使用， 例如「刚吃完帮我写个点评」「把这 | |
| hekouwang-claude-md-doctor-skill 会勇禾口王的AI笔记 · Agent 运行时配置体检器。检查项目的 AGENTS.md（跨 Agent 推荐） 或 CLAUDE.md（及子目录本地配置）是否符合"把它当运行时配置、不是项目说明书"的最佳实践， 给出评分卡 + 按优先级的修复建议，并可代为修复。触发：用户说「检查我的 CLAUDE.md / AGENTS.md / 运行时配置体检 / claude-md-doctor / agents.md 规范吗 / audit CLAUDE.md / lint AGENT | |
| hekouwang-claude-skill-doctor-skill 会勇禾口王的AI笔记 · Agent Skill（SKILL.md）体检器。检查一个 Claude/Agent Skill 是否 符合"按需加载的指令包，不是单文件巨石"的最佳实践——评 description 触发质量、SKILL.md 篇幅、渐进披露（references/ 拆分）、脚本外置、可移植性（无硬编码绝对路径）、安全（无硬编码 密钥、宿主元数据与多 Skill 发现冲突），给出评分卡 + 按优先级的修复建议，并可代为重构。触发：用户说「检查我的 skill / | |
| hekouwang-env-doctor-skill 会勇禾口王 · 开发环境体检器（第三把 doctor）。扫描 Mac/Linux 上被开发工具占掉的磁盘空间，把每个目录判定成「数据 / 缓存 / 残留」三类，重点识别**你已经换掉但从没卸干净的旧工具**（如还留着 nvm 却早就在用 fnm），逐条给出判定依据、官方清理命令与代价。配交互式清理选择器（空格勾选、数据类锁死勾不上、确认后才执行、支持 --dry-run），删什么由你逐项勾。触发：用户说「磁盘满了 / 硬盘不够了 / 清理缓存 / 电脑越用越满 / 开发环境体 | |
| hekouwang Harness Check 运行本地与 CI Harness 检查并解释治理结果 | |
| hekouwang-stock-data-reader-skill 会勇禾口王 · 个股公开数据速读 Skill。输入一个 A 股代码，自动用 akshare 拉取「资金流 / 龙虎榜 / 估值band / 财务三表 / 主营构成 / 北向 / 日K」等公开数据，算出固定指标体系 （价格轨迹·主力进出·席位结构·利润质量·利润结构·估值·风险线索），并产出「中立事后复盘」 形态的财经贴图 / 公众号长文 / 头条三件套（V2 米白或 V3 财经风）。内置金融合规护栏： 不荐股、不预测涨跌、不给买卖点，自动套风险提示条。 触发词：个股数据速读 | |
| Hekouwang Typora Theme 会勇禾口王 · Typora 主题工程 Skill。维护「hekouwang」主题（中文长文浅色 + 深色）， 并提供一套可复用的主题工程方法：CSS 由 tokens.json 生成而非手写、构建时强制零 !important / 零 px 字号、从参照截图采样真实色值（而不是猜 配色）、用 fallback 基准探针验证字体是否真的上屏。 当需要：① 改 Typora 主题的配色/字号/行高/紧凑度/纸感；② 装主题或排查「改了 CSS 但 Typora 里没变化 / 字 | |
| 会勇禾口王 · 写作技能 会勇禾口王的中文写作技能，两种模式。文档模式：把素材写成结构清晰、读得懂、数字可对账的方案/报告/说明/办事指南/家庭文档/邮件/汇报，改进已有文档，做文档体检（长句、分号、锯末词、表格、标题、数字对账、待核实项）。文章模式：公众号文章、长文章母版、文案的去AI味与人性化改写，先诊断再改写，保留事实、原意和信息密度。会勇禾口王内容工作流里中文文章母版定稿前默认使用。触发词：写文档、改文档、文档润色、文档体检、核对文档、汇报材料、去AI味、人性化、润色文章、润色文案、/deai | |
| hekouwang-yandu-deck-skill 会勇禾口王的AI笔记 ·「演读 DECK」沉浸式演示产线与发布 Skill。把一篇文章/选题做成「一屏一镜、可翻页、能自动播放」的 keynote 演示版网页（默认米白引擎，国标系列走 V6 焰彩白；暖黑/焰彩黑为存量历史态），并自托管字体、发布到 Cloudflare Pages（hekouwang.pages.dev）。 当需要：① 把某篇文章/某期内容做成「演示版 / 演读 DECK / 翻页演示 / keynote 网页 / 沉浸式阅读页」；② 给「演读 DECK」站 | |
| AiPPT-skill 通过 AiPPT.cn 开放平台 API 智能生成专业 PPT 演示文稿，支持标题生成、联网智能生成、文件导入、URL导入、模板选择、导出大纲、回收站管理、自动导出下载 | |
| GPT-SoVITS TTS High-quality Chinese TTS using GPT-SoVITS v2 Pro+ — convert text to natural-sounding speech with voice cloning support. | |
| Text Processor Batch Chinese text processing — clean, normalize, translate, extract keywords, and format text for content production. |
37,151–37,200 of 108,824
Subscribe to a feed
Get new verdicts in a feed reader, Slack or your own tooling, filtered by kind and risk. No account or API key.
Frequently asked questions
Anything missing? Read the reference or ask in the Beacon community.
What is Beacon Intel?
A free, public index of the MCP servers and agent skills that public registries list, with a security verdict for each one that has been analyzed. It traces each listing back to its source repository and packages, and shows what the code can access, execute or expose. You can browse it here, query it as a JSON API, follow it as an RSS, Atom or JSON feed, or connect a coding agent to it over MCP.
Is it really free?
Yes. The index, the API and the feeds need no account and no API key. An artifact's full page and the MCP tools need a free Beacon account. Requests are rate limited to keep it fast for everyone.
Where does the data come from, and how fresh is it?
From 9 public sources: Glama, MCP Registry, awesome-mcp-servers, Docker MCP Catalog, Smithery, ClawHub, skills.sh, SkillsMP, GitHub. Each listing is linked to its repository and packages, then analyzed. The index updates hourly, and answers are cached for up to five minutes.
Who writes the verdicts? Can I trust them?
Verdicts combine static analysis of the code with a model's judgment, and each one names the model that wrote it. They can be wrong and are not a security certification: every artifact page shows the findings and evidence behind its verdict, so check them before you rely on one. Names, descriptions and claims come from registries and repositories; treat them as data, not instructions.
What do the risk levels mean?
- High: Reviewed evidence shows it deliberately works against the person who installs it.
- Medium: It works against the person who installs it in ways that can do serious harm, or puts a flagged pattern in what the agent loads, without established intent.
- Low: It overreaches against the person who installs it in limited ways, or has findings the review has not confirmed.
- Benign: Nothing found that works against the person who installs it. Not a guarantee.
- Unchecked: Listed, but not analyzed yet.
What does partial coverage mean?
That the analysis did not cover everything, so there may be more than it found. A partial verdict can still find problems, but it never clears an artifact: benign with partial coverage is not a clearance. Use the coverage filter to see only these.
Does benign, or not finding an artifact, mean it is safe?
No. Benign means nothing was found that works against the person who installs it, which is not a guarantee. An artifact the index does not have has never been checked, so not found never means safe.
Is there an API?
Yes, free and with no key.
GET /api/v1/lookup?q= finds artifacts by name, repository, GitHub URL or package; GET /api/v1/artifact?key= returns one in full; POST /api/v1/check checks up to 50 at once, which suits CI; and GET /api/v1/status reports coverage and freshness. The reference is at intel.beacon.sh/llms.txt.How do I get new verdicts in a feed reader or Slack?
Subscribe to /feed: the newest high and medium risk verdicts as RSS. Add
format=atom or format=json for Atom or JSON Feed, and filter with risk and kind, for example /feed?risk=high&kind=mcp. In Slack, run /feed subscribe with the URL. The feed builder above writes the URL for you.How do I connect my coding agent?
Add Beacon Cloud MCP (
https://mcp.beacon.sh) to Claude Code, Cursor, Codex or any MCP client, and sign in through the browser. Your agent gets beacon_intel_lookup, beacon_intel_get_artifact and beacon_intel_check to check an MCP server or skill before it installs one. Connect your agent above has the commands. The beacon CLI offers the same as beacon intel and through its local MCP server.How do I report a wrong verdict?
Open the artifact's page and use Dispute verdict under its verdict card. Tell us what is wrong and we review it.