Browser Act

记录自身发现以实现自我改进的技能

用于 AI 代理的浏览器自动化 CLI。当用户需要与网站交互（包括浏览页面、填写表单、点击按钮、截图等）时使用。

自我反思+自我批评+自我学习+自组织记忆。智能体评估自身工作、发现错误并持续改进。

AI智能体技能安全预审工具。安装ClawdHub、GitHub等来源技能前，检查风险信号、权限范围及可疑模式。

类型化知识图谱，用于结构化智能体记忆与可组合技能。适用于以下场景：创建/查询实体（人物、项目、任务、事件、文档）、关联相关对象、强制执行约束、将多步操作规划为图谱变换，或当技能需要共享状态时。触发关键词包括"记住""我知道关于什么""将X链接到Y""显示依赖关系"、实体增删改查操作，以及跨技能数据访问。

将 AI 代理从任务执行者转变为能够预见需求、持续优化的主动伙伴。现已支持 WAL Protocol、Working Buffer、Autonomous Crons 及经过实战验证的模式。属于 Hal Stack 🦞。

专为AI智能体优化的无头浏览器自动化CLI，支持无障碍树快照和基于引用的元素选择。

创建有效技能指南。当用户希望创建新技能（或更新现有技能）以利用专业知识、工作流程或工具集成扩展 Claude 的能力时，应使用此技能。

从网站探索锻造可复用 Skill 包（SKILL.md + 脚本），无需再探索。适用于用户想要可复用 Skill。

Fetch Xiaohongshu (RedNote / xhs) user profile information and their published notes list by user ID, returning nickname, bio, follower/following counts, engagement totals, tags, and paginated notes with engagement stats. Use when user mentions user profile xiaohongshu, rednote creator profile, xhs influencer data, scrape xiaohongshu user, get blogger notes, xiaohongshu author info, KOL discovery xiaohongshu, rednote user stats, creator profile rednote, xiaohongshu blogger analysis, get xhs user followers, rednote influencer profile, xiaohongshu account info, xhs creator data, user notes list xiaohongshu, rednote account scrape, xiaohongshu KOL research.

Scrape job listings from Indeed.com by keyword, location, and country. Returns job title, company, salary, rating, description, benefits, and apply links. Use when user mentions Indeed, Indeed scraper, Indeed jobs, scrape Indeed, job search Indeed, Indeed job listings, extract Indeed data, Indeed job data, get jobs from Indeed, Indeed employment data, job market research Indeed, Indeed salary data, bulk job extraction Indeed, Indeed job scraper, monitor Indeed listings, Indeed hiring data, job postings Indeed, Indeed career search. Also applies to: job market analysis, salary benchmarking from Indeed, competitor hiring monitoring, recruitment data collection, building job databases from Indeed search results.

Scrapes posts from a Facebook group given a group URL, sort order, and desired count — returns structured post metadata including post_id, permalink, author, timestamp, body text, images/videos, reaction counts, reaction type breakdown, comment count, and share count. Use when: user wants to scrape/

X (Twitter) DM automated chat end-to-end Skill: scan DM inbox to identify pending-reply conversations, read message history, generate persona-based replies and send; also supports searching users and starting new conversations. Built-in E2E passcode unlock, DM permission filtering, and rate control. Use when user mentions X auto-reply DMs, Twitter DM automated chat, auto-handle unread DMs, reply to X private messages with persona, X DM outreach campaign, batch send DMs to Twitter users, auto-process pending DM replies, Twitter DM bot, automated Twitter outreach, X direct message automation.