0xmariowu0xmariowuCommunityΒ·63 Agent Skills Included

Autosearch

Multi-channel deep research across 40 global and Chinese sources

Searches 40 channels at once, including arxiv, GitHub, Reddit, YouTube, Xiaohongshu, Weibo, and Zhihu, then deduplicates and ranks every result with a source URL. Eliminates manual cross-platform research, training-data cutoffs, and shallow single-engine web results. Routes each query to the right channels, transcribes video and audio, and learns from past sessions to improve future searches.
npx skills add 0xmariowu/Autosearch --all -g -y
Available:

Instructs the agent to run every configured channel per search, score results by relevance instead of popularity, reflect after each round, and record winning patterns so the system improves across sessions.

All Skills in This Repository (63)

Pure Emerald Level Indicators
πŸ“¦ In Repo
0xmariowu0xmariowu

valid_channel

Search bilingual scholarly papers and hydrate selected results with detailed metadata.

Community
Intermediate
πŸ“¦ In Repo
0xmariowu0xmariowu

valid_tool

Fetch JSON data from HTTP endpoints using curl-based requests.

Community
Basic
πŸ“¦ In Repo
0xmariowu0xmariowu

bad_fallback

Validate fallback_chain method ids in skill metadata against defined methods.

Community
Intermediate
πŸ“¦ In Repo
0xmariowu0xmariowu

bad_requires_token

Validate requires tokens in SKILL.md frontmatter and reject malformed entries.

Community
Basic
πŸ“¦ In Repo
0xmariowu0xmariowu

stub_ok

Echo input queries as deterministic Evidence objects with metadata.

Community
Basic
πŸ“¦ In Repo
0xmariowu0xmariowu

stub_cookie

Expose a single mock method enforcing cookie-based access control.

Community
Basic
πŸ“¦ In Repo
0xmariowu0xmariowu

autosearch:router

Route autosearch tasks to 1–3 skill groups and 3–8 leaf skills.

Community
Advanced
πŸ“¦ In Repo
0xmariowu0xmariowu

fetch-crawl4ai

Fetch dynamic pages and render JavaScript-heavy content with Playwright-backed crawling.

Community
Advanced
πŸ“¦ In Repo
0xmariowu0xmariowu

video-to-text-openai

Transcribe video or audio into text and SRT subtitles using yt-dlp and OpenAI Whisper.

Community
Advanced
πŸ“¦ In Repo
0xmariowu0xmariowu

video-to-text-local

Transcribe YouTube URLs and local media to text and SRT on Apple Silicon.

Community
Advanced
πŸ“¦ In Repo
0xmariowu0xmariowu

fetch-playwright

Drive a Chromium browser via the Playwright MCP server for dynamic web interactions.

Community
Intermediate
πŸ“¦ In Repo
0xmariowu0xmariowu

mcporter

Route free MCP servers for Exa, Weibo, Douyin, Xiaohongshu, and LinkedIn through mcporter.

Community
Basic

Frequently Asked Questions

FAQPage Schema
How to install AutoSearch?β–Ό

Run `npx skills add 0xmariowu/Autosearch --all -g -y` in your terminal to install all skills globally. Afterward, run `autosearch doctor` to verify which channels are active.

What sources can AutoSearch search?β–Ό

It covers 40 channels including arxiv, PubMed, GitHub, Reddit, Hacker News, YouTube, Stack Overflow, and 10+ Chinese platforms like Xiaohongshu, Weibo, Zhihu, Bilibili, and Douyin. Many work with no API key; a few need a free or paid key.

Does AutoSearch work with Claude Code and Cursor?β–Ό

Yes. It is MCP-native and plugs into Claude Code, Cursor, Zed, Cline, LangChain, and other hosts without changing your model or prompts.

Can AutoSearch transcribe videos and podcasts?β–Ό

Yes. Built-in skills transcribe YouTube, Bilibili, Douyin, and local media files to text and subtitles using free Groq or local Whisper, with paid OpenAI as a fallback.

Do I need coding skills to use AutoSearch?β–Ό

No. Once installed, you just ask your agent a research question in plain language and it picks channels, runs searches, and returns cited results automatically.

Related Repositories in Data & Analytics

View All in Data & Analytics→