All authors
Claude Skills by 0xmariowu
github.com/0xmariowu70 skills1 installs87 views
- Reflective Search LoopExplicit loop state for multi-round research — maintains gaps, visited URLs, bad URLs, failed-evaluator feedback, and decides whether to fire another round of queries or stop. Inspired by WebThinker's in-band search protocol, node-deepresearch's agent.ts state, and Scira's extreme-search.Votes: 0GitHub stars: 44
- Tikhub FallbackDecision tree for when to escalate from free native Chinese channels (bilibili / weibo / xiaohongshu / douyin / zhihu) to the paid TikHub fallback. ~$0.0036 per request; 5 platforms covered. Tells the runtime AI when the cost is worth paying vs. when the free paths are enough.Votes: 0GitHub stars: 44
- Trace HarvestDistill reusable knowledge from successful session tool-call traces. Reads runtime tool-call logs + Evidence outputs + user acceptance signals, filters successful paths, and writes compact patterns to per-skill experience/patterns.jsonl. Differs from outcome-tracker (which only records downstream acceptance) by parsing the full trace structure — inputs, tool choices, branches, recoveries.Votes: 0GitHub stars: 44
- RouterRoute an autosearch task to the smallest relevant group of skills, then load only the matching leaf skills. First line of defense against loading all 80+ skill SKILL.md files at session start.Votes: 0GitHub stars: 44
- Fetch Crawl4aiDeep URL fetch using crawl4ai (Playwright-powered) for JS-rendered pages, anti-bot sites, and dynamic content. Slower than fetch-jina but handles sites that block simple fetchers. Requires user-installed crawl4ai package.Votes: 0GitHub stars: 44
- Fetch FirecrawlUse to scrape a URL into clean Markdown when the page is JS-rendered, behind anti-bot, or a PDF — and FIRECRAWL_API_KEY is available. Degrades gracefully to warn when key is missing.Votes: 0GitHub stars: 44
- Fetch JinaFetch any public URL as clean Markdown via Jina Reader (r.jina.ai). Best for articles, docs, blog posts; weak on heavy anti-bot sites (Zhihu, Xiaohongshu).Votes: 0GitHub stars: 44
- Fetch PlaywrightInteractive browser automation (click, type, wait, screenshot, navigate) for dynamic pages that need user-like actions — pagination, login, form submission, infinite scroll. Uses Microsoft's official @playwright/mcp server; no autosearch-side Python. Use when fetch-jina and fetch-crawl4ai still cannot get the content.Votes: 0GitHub stars: 44
- McporterRoute the runtime AI to free third-party MCP servers via the upstream `mcporter` router — semantic web search (Exa), Chinese UGC (Weibo / Douyin / Xiaohongshu), and professional (LinkedIn). Zero API keys; free fallback path before reaching for paid TikHub on hard Chinese anti-bot platforms.Votes: 0GitHub stars: 44
- Video To Text BcutTranscribe video/audio URL to text + word-level timestamps using Bilibili Bcut ASR API (free, no API key). Preferred for Chinese content — Bcut gives character-level timestamps vs Whisper word-level. Returns text + segments [{start, end, text}]. Requires yt-dlp + ffmpeg.Votes: 0GitHub stars: 44
- Video To Text GroqTranscribe video/audio URL or local file to text + SRT using yt-dlp + Groq Whisper API (free tier). Preferred default for v2 transcription. Returns raw text and segments; summary is caller's responsibility.Votes: 0GitHub stars: 44
- Video To Text LocalTranscribe video/audio URL or local file to text + SRT using yt-dlp + local mlx-whisper (Apple Silicon). Free, offline, fastest on M-series Macs. Opt-in advanced path for users with Apple Silicon + mlx-whisper installed. Returns raw text and segments; summary is caller's responsibility.Votes: 0GitHub stars: 44
- Video To Text OpenaiTranscribe video/audio URL or local file to text + SRT using yt-dlp + OpenAI Whisper API. Paid fallback for v2 transcription when Groq is rate-limited or unavailable. Returns raw text and segments; summary is caller's responsibility.Votes: 0GitHub stars: 44
- Bad Fallback<!-- Self-written, plan autosearch-0418-channels-and-skills.md § F001 --> --- name: bad_fallback description: Contains a fallback reference that does not exist. version: 1 methods: - id: primary impl: impl.py fallback_chain: [primary, secondary] --- The loader should reject unknown fallback method ids.Votes: 0GitHub stars: 44
- Bad Requires Token<!-- Self-written, plan autosearch-0418-channels-and-skills.md § F001 --> --- name: bad_requires_token description: Contains an invalid requires token on purpose. version: 1 methods: - id: api impl: impl.py requires: [cookie:valid_cookie, invalid-token] fallback_chain: [api] --- The loader should reject malformed `requires` entries.Votes: 0GitHub stars: 44
- Stub Cookie--- name: stub_cookie description: "Stub channel whose only method requires a cookie." version: 1 languages: [en] methods: - id: fake impl: methods/fake.py requires: [cookie:stub_cookie] fallback_chain: [fake] when_to_use: query_languages: [en] query_types: [experience-report] avoid_for: [] quality_hint: typical_yield: low chinese_native: false ---Votes: 0GitHub stars: 44
- Stub Ok--- name: stub_ok description: "Stub channel with one always-available echo method." version: 1 languages: [en] methods: - id: echo impl: methods/echo.py requires: [] fallback_chain: [echo] when_to_use: query_languages: [en] query_types: [technical] avoid_for: [] quality_hint: typical_yield: low chinese_native: false ---Votes: 0GitHub stars: 44
- Missing Required<!-- Self-written, plan autosearch-0418-channels-and-skills.md § F001 --> --- description: Missing the required name field on purpose. version: 1 methods: - id: primary impl: impl.py fallback_chain: [primary] --- The loader should reject this frontmatter because `name` is required.Votes: 0GitHub stars: 44
- Valid Channel<!-- Self-written, plan autosearch-0418-channels-and-skills.md § F001 --> --- name: valid_channel description: Use for bilingual paper-style searches that may need a detail lookup. version: 1 languages: [zh, en] methods: - id: api_search impl: methods/api_search.py requires: [env:ARXIV_TOKEN] rate_limit: per_min: 30 per_hour: 500 - id: api_detail impl: methods/api_detail.py requires: [cookie:arxiv_session] rate_limit: per_min: 10 fallback_chain: [api_search, api_detail] when_to_use: query_lan...Votes: 0GitHub stars: 44
- Valid Tool<!-- Self-written, plan autosearch-0418-channels-and-skills.md § F001 --> --- name: valid_tool description: Use for reusable HTTP fetch helpers shared across channels. version: 1 methods: - id: fetch_json impl: impl.py requires: [binary:curl] fallback_chain: [fetch_json] quality_hint: typical_yield: unknown chinese_native: false --- This tool skill exposes a single primitive and does not need routing hints.Votes: 0GitHub stars: 44