القائمة
Searches the web using the built-in web_search tool. Use when the user asks for fresh information, web research, recent news, documentation lookup, or current facts.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Generates images, videos, and other assets using built-in tools. Use when the user asks to create or edit images, generate videos, or produce other media assets.
Releases JAcoworks Desktop (Tauri) — version bump, macOS/Windows build, COS upload, DB registration, git tag. Use when the user says 'release', 'publish', '发版', '发布版本', or mentions a version number like 'v1.6.1'.
Automates web pages through the bundled agent-browser CLI. Use when the user asks to open a page, click or fill elements, take screenshots, inspect DOM state, reuse Chrome login sessions, or perform browser-based workflows via bash.
创建和管理 Agent 技能包。 创建和管理 Agent 技能包。当用户要求创建技能、新建 skill、编写 SKILL.md 时使用。Triggers on: 创建技能, 新建skill, create skill, build skill.
Create and export 3D scenes using Blender 4.4 bpy scripts + Kenney GLB assets for Three.js. Use when building office environments, assembling furniture layouts, or exporting GLB for the observatory.
OpenClaw container integration with JAcoworks gateway. Use when deploying, configuring, debugging, or extending OpenClaw containers, the WS protocol bridge, or container_type dispatch logic.
| name | ai-search |
| description | Searches the web using the built-in web_search tool. Use when the user asks for fresh information, web research, recent news, documentation lookup, or current facts. |
Use the built-in web_search tool directly — do NOT use ai-search CLI through bash.
web_search(query="latest Bun release notes")
web_search(query="how does bun build --compile work", mode="answer")
web_search(query="vector database benchmark 2026", num=8)
| Mode | Use when |
|---|---|
fast | Default. Quick factual lookups, docs, news |
answer | Need a synthesized summary or direct answer |
ai-search CLI through bash — use web_search tool directlydeep — it is slow and rarely neededcontent-extract skill insteadbrowser skill instead