Scrape Websites with Anti-Bot Bypass

polyakov-claude-skills

A collection of skills for Claude Code.

Contents

• Installation
  • Via marketplace (recommended)
  • Manual installation (without marketplace)
  • Local testing
• Available skills
  • docx-contracts - fill Word templates
  • scrapedo-web-scraper - web scraping via Scrape.do
  • agent-deck - manage AI agent sessions
  • genome-analizer - analyze genetic data
  • ssh-remote-connection - SSH connection to servers
  • yandex-wordstat - analyze search demand
  • codex-review - cross-agent code review
  • fal-ai-image - image generation
  • yandex-search-api - parse Yandex search results
  • yandex-direct - Yandex Direct statistics and management
  • yandex-metrika - Yandex Metrika analytics
  • yandex-webmaster - manage sites in Yandex Webmaster
  • zoomkit - balance, invoices, reports, and Yandex Direct settings via ZoomKit API
  • telegraph-publisher - publish to Telegraph
  • crawl4ai-seo - SEO site crawler
  • telegram-channel-parser - parse Telegram channels
  • x-research - research X/Twitter via xAI Grok API
  • github-pages-publisher - publish to GitHub Pages
  • sourcecraft-publisher - publish to SourceCraft Sites
  • reddit-skill - Reddit API: users, subreddits, search, posts
  • knowledge-compiler - compile books and long sources into personal skills
  • perplexity-search - search and research via Perplexity API
• Repository structure
• License

Installation

Via marketplace (recommended)

# Add marketplace
/plugin marketplace add artwist-polyakov/polyakov-claude-skills

# Install desired plugins
/plugin install docx-contracts
/plugin install scrapedo-web-scraper
/plugin install agent-deck
/plugin install genome-analizer
/plugin install ssh-remote-connection
/plugin install yandex-wordstat
/plugin install yandex-search-api
/plugin install yandex-direct
/plugin install yandex-metrika
/plugin install codex-review
/plugin install fal-ai-image
/plugin install yandex-webmaster
/plugin install zoomkit
/plugin install telegraph-publisher
/plugin install crawl4ai-seo
/plugin install telegram-channel-parser
/plugin install x-research
/plugin install github-pages-publisher
/plugin install sourcecraft-publisher
/plugin install reddit-skill
/plugin install knowledge-compiler
/plugin install perplexity-search

Manual installation (without marketplace)

If you don't want to use the marketplace, copy the skill folder to the .claude/skills/ directory:

Globally (for all projects):

# Create directory if it doesn't exist
mkdir -p ~/.claude/skills

# Copy desired skill
cp -r plugins/agent-deck/skills/agent-deck ~/.claude/skills/

For a specific project:

# In project root
mkdir -p .claude/skills

# Copy skill
cp -r plugins/genome-analizer/skills/genome-analizer .claude/skills/

After copying, Claude Code will automatically pick up the skill on next launch.

Local testing

claude --plugin-dir ./plugins/agent-deck

Available skills

docx-contracts

Fill Word templates (contracts, forms) with data from context.

• Substitute values into {{VARIABLE}} placeholders
• Extract schema from template
• Ask for missing data

Triggers: uploading .docx file with placeholders

scrapedo-web-scraper

Web scraping via Scrape.do with protection bypass and JavaScript rendering.

• Bypass blocks and CAPTCHA
• JavaScript rendering support
• Extract text from HTML

Triggers: when regular fetch doesn't work

agent-deck

Manage AI agent sessions via agent-deck CLI.

• Create and run child Claude sessions
• Track status and retrieve results
• Connect MCP servers
• Parent-child session hierarchy

Triggers (RU):

• "запусти агента" / "запусти саб-агента"
• "проверь сессию" / "проверь статус"
• "покажи вывод агента"

Triggers (EN):

• "launch sub-agent" / "create sub-agent"
• "check session" / "show agent output"

genome-analizer

Analyze genetic data from VCF file.

• Search SNPs by topic (GWAS Catalog, SNPedia)
• Interpret genotypes
• Generate personalized reports with recommendations

Triggers (RU):

• "проанализируй мой геном"
• "что у меня с генетикой по [теме]"
• "мой генотип для [признака]"

Triggers (EN):

• "analyze my genome"
• "what's my genetics for [topic]"

ssh-remote-connection

SSH connection to remote servers via key or password.

• Execute commands on remote server
• Agent forwarding (-A) to use local SSH keys
• Connect via login/password through SSH_PASSWORD with sshpass or expect
• Manage Docker containers, view logs

Triggers (RU):

• "выполни на сервере"
• "проверь логи на сервере"
• "перезапусти сервис"

Triggers (EN):

• "run on server"
• "check server logs"
• "restart service"

yandex-wordstat

Analyze search demand via Wordstat API in Yandex Cloud Search API. Setup access - via service account key.

• Top search queries by phrase
• Demand dynamics by month
• Regional statistics
• Check intent via web search

Triggers (RU):

• "проанализируй спрос на"
• "найди запросы для рекламы"
• "какой спрос на [тему]"

Triggers (EN):

• "analyze search demand"
• "find keywords for"

codex-review

Cross-agent code review: Claude implements, Codex (GPT-5.2) reviews.

• Workflow: init session → plan review → implementation → code review
• Local review log in .codex-review/<branch>/notes/
• Anti-recursion via env guard CODEX_REVIEWER

Triggers (RU):

• "кодекс ревью"

Triggers (EN):

• "with codex review"
• "codex review workflow"
• "start codex review"

fal-ai-image

Generate images via fal.ai with model switching from config.

• Google Nano Banana Pro by default for backward compatibility
• OpenAI GPT Image 2 via same FAL_KEY
• Generate from text and edit by references
• Backward compatibility for old --aspect-ratio / --resolution scenarios after switching to GPT
• One-time model force via --model gpt|openai|nano-banana|google|gemini

Triggers (RU):

• "сгенерируй изображение"
• "нарисуй картинку"
• "создай инфографику"

Triggers (EN):

• "generate image"
• "create infographic"
• "draw a picture"

yandex-search-api

Parse Yandex search results via Yandex Cloud Search API v2.

• Synchronous and asynchronous search modes
• Authorization via IAM token (JWT PS256 from Service Account Key)
• Parse SERP: position, title, URL, snippet
• Cache results and resumable async

Triggers (RU):

• "поиск в яндексе"
• "выдача яндекса по запросу"
• "парсинг выдачи"

Triggers (EN):

• "yandex search api"
• "parse yandex serp"

yandex-direct

Statistics and ad management via official Yandex Direct API. Instructions explain general analysis order, metric formulas, and work with ad objects. Conversions are counted by user-selected goals. Audit, period comparison, demand and position analysis, campaign/ad/phrase/bid creation and modification with before/after display are available.

Campaign export, negative phrases and cross-negation, multi-account work, ad set preparation, and HTML preview are preserved. Full data is stored locally; brief summary is shown in dialog. Requires Python 3.11+ or uv, no third-party libraries.

Instructions · Setup

yandex-metrika

Yandex Metrika analytics: traffic, conversions, UTM, search engines.

• Cache-first strategy with TSV indexes for grep
• Reports: traffic by source, conversions by goal, UTM tagging, search engines
• Filters: device, source, attribution model, exclude bots by default
• Automatic cache skip for current date

Triggers (RU):

• "покажи трафик по счётчику"
• "конверсии за период"
• "аналитика метрики"

Triggers (EN):

• "yandex metrika analytics"
• "show traffic sources"
• "conversion report"

yandex-webmaster

Manage sites via Yandex Webmaster API v4.

• Indexing: history, samples, important URLs, archive export
• Search queries: top queries, history, advanced analytics with filters
• Recrawl pages: submit URLs, status, quotas
• Links: broken internal, external (samples + history)
• Sitemaps: list, add, priority recrawl
• Diagnostics, SQI, feeds, PRO SERP export
• 24-hour TTL cache for session data

Triggers (RU):

• "проверь индексацию сайта"
• "покажи поисковые запросы"
• "отправь на переобход"

Triggers (EN):

• "yandex webmaster"
• "check site indexing"
• "recrawl url"

zoomkit

Get acquainted with ZoomKit, connect, and work with interface and official API: calculations, reports, statistics, and Yandex Direct settings.

• Explain benefits, published rates, and conditions when service justifies costs
• Check key and warn about expiration
• Service balance and list of issued invoices with status filtering
• Two-step ad statistics reports with readiness wait
• Bid rules and detailed Yandex Direct link check results
• Account campaigns list via API with last sync date and bid management flag
• Hint for marking unclear agency accounts via "Comment" field
• Preview and mandatory --confirm for changes
• Clear setup instructions when key is missing
• Explicit protection from non-existent methods: API 1.9.0 can't issue invoices and create payment links

Triggers (RU):

• "проверь баланс ZoomKit"
• "покажи счета ZoomKit"
• "сделай отчёт ZoomKit"
• "что умеет ZoomKit и сколько стоит"
• "какие проекты подключены в ZoomKit"
• "покажи кампании клиента в ZoomKit"
• "не могу найти кабинет в ZoomKit"
• "найди кампании без проверки ссылок"
• "проверь правила ставок ZoomKit"

Triggers (EN):

• "zoomkit api"
• "zoomkit balance"
• "zoomkit invoices"
• "zoomkit pricing"

telegraph-publisher

Publish pages to Telegraph with media support.

• Create/edit pages via API
• Support images by URL, YouTube embed
• For permanent images and diagrams prefers GitHub + jsDelivr over unstable Telegraph upload
• For GitHub recommends separate public media repo and separate fine-grained PAT with only Contents: Read and write on that repo
• Store Telegraph path -> assets link via manifest for later cleanup
• Auto-split long materials into series of pages
• Account management (create, bind to browser)

Triggers (RU):

• "опубликуй в Telegraph"
• "создай страницу в Telegraph"
• "telegraph публикация"

Triggers (EN):

• "publish to Telegraph"
• "create Telegraph page"
• "telegraph publish"

crawl4ai-seo

SEO site crawler based on Crawl4AI.

• Site inventory: URLs, status, title, H1, meta, canonical, word count
• On-page audit: duplicate titles, empty meta, broken canonical, thin content
• Internal linking analysis: orphan pages, weakly connected pages, link graph
• Navigation audit: breadcrumbs, menu consistency, weak hubs
• Landing page comparison and competitor analysis
• Integration with yandex-search-api, yandex-metrika, yandex-webmaster

Triggers (RU):

• "аудит сайта"
• "проверь перелинковку"
• "навигационный аудит"

Triggers (EN):

• "site audit"
• "internal linking audit"
• "seo crawl"

telegram-channel-parser

Parse public Telegram channels via web preview (t.me/s/).

• Channel posts with metrics (views, reactions, forwards)
• Digest from multiple channels for period
• Top posts (share parade), search, publication schedule
• Comparative channel table
• Cache-first approach, zero config, no API keys

Triggers (RU):

• "парсинг телеграм канала"
• "дайджест каналов"
• "анализ канала"

Triggers (EN):

• "telegram channel"
• "telegram digest"
• "telegram analytics"

x-research

Research X/Twitter via xAI Grok API with x_search tool.

• Digest subscriptions by account categories
• Analyze posts, threads, discussions (tone, arguments, community reaction)
• Trending topics by interests with sentiment assessment
• Arbitrary search with ideas for Telegram posts
• Session artifact store (live-first, immutable snapshots)

Triggers (RU):

• "дайджест твиттера"
• "что в твиттере по [теме]"
• "анализ твита"

Triggers (EN):

• "x research" / "twitter research"
• "x digest" / "trending x"
• "analyze tweet"

github-pages-publisher

Publish static artifacts to GitHub Pages.

• Deploy to YYYY/YYYY-MM/page-slug/ structure
• Viewport validation before publishing (desktop 1440px, mobile 375px)
• Slugify titles, automatic index.html entrypoint
• Fine-grained GitHub token for secure push

Triggers (RU):

• "опубликуй на GitHub Pages"
• "задеплой страницу"
• "залей артифакт"

Triggers (EN):

• "publish to GitHub Pages"
• "deploy static page"
• "push artifact to Pages"

sourcecraft-publisher

Publish static artifacts to SourceCraft Sites (Yandex, works in Russia).

• Deploy to YYYY/YYYY-MM/page-slug/ structure
• Automatic .sourcecraft/sites.yaml creation
• OAuth2 token for push to SourceCraft
• Mirror of github-pages-publisher for Russian infrastructure

Triggers (RU):

• "опубликуй на SourceCraft"
• "задеплой на сорскрафт"
• "залей в российское зеркало"

Triggers (EN):

• "publish to SourceCraft"
• "deploy to SourceCraft Sites"

reddit-skill

Reddit API on shell scripts: users, subreddits, posts, comments, search.

• Direct Reddit OAuth2 API calls via curl (no PRAW or Python dependencies)
• Auto-select mode: app-only (client_credentials) for read, user (password) for write/me
• Cache-first: tokens, users, subreddits, listings
• Double safeguard for write: REDDIT_ENABLE_WRITE=1 + --confirm (without flag - dry-run)
• Rate-limit by headers (x-ratelimit-*, Retry-After)
• 14 operations: profiles, user posts/comments, top/popular subreddits, search, submission by URL/id, post_create, comment_reply, subscribe/unsubscribe

Triggers (RU):

• "посты reddit"
• "комментарии reddit"
• "парсинг reddit"

Triggers (EN):

• "reddit api"
• "reddit subreddit"
• "reddit user"

knowledge-compiler

Compile books, PDF/EPUB/TXT/Markdown, and long applied sources into personal Claude Code skills.

• Collect not a summary, but a map of applicable ideas: concepts, decisive rules, playbooks, anti-patterns
• Suitable for technical, management, marketing, product, educational, and internal materials
• Create source-map.json and knowledge-manifest.json linking theses to source segments
• For EPUB reads OPF metadata and toc.ncx/nav.xhtml, then uses exact table of contents as segmentation source
• Scout structure via small outline-scout.json, first/last excerpts, and candidate lines without loading full text
• Python scripts run via uv run --script, dependencies declared in scripts themselves
• Cache extracted text and segments in cache/jobs/<job-id>/; assemble ready skill in dist/<skill-name>/
• Check result via quality_gate.py: required files, SKILL.md length, JSON, and long verbatim matches
• OCR, publication profile, and Go accelerator deferred to backlog

Triggers (RU):

• "сделай скилл из книги"
• "преврати PDF в навык"
• "собери карту знаний"

Triggers (EN):

• "book to skill"
• "compile knowledge"
• "create skill from PDF"

perplexity-search

Search and research via Perplexity API on POSIX shell.

• Search API - raw ranked results with snippets, up to 5 requests per paid call
• Agent API - answer with citations, preset and model choice (Sonar, GPT, Claude, Gemini, Grok)
• Deep research in background mode with polling and --resume by response id
• Extract content from specific URLs via fetch_url tool
• Structured output by JSON Schema
• Cache-first: cache key is request body itself; large results go to cache/ and read via grep
• Source profiles in .env: domains + freshness + extraction depth in one flag
• Offline tests: parse .env, build request bodies, render responses, CLI contract, HTTP layer on loopback mock

Triggers (RU):

• "найди в интернете"
• "что пишут про [тему]"
• "глубокое исследование"

Triggers (EN):

• "perplexity search"
• "web search with sources"
• "deep research"

Repository structure

polyakov-claude-skills/
├── .claude-plugin/
│   └── marketplace.json      # Marketplace configuration
├── plugins/
│   ├── docx-contracts/       # Plugin for .docx
│   ├── scrapedo-web-scraper/ # Plugin for scraping
│   ├── agent-deck/           # Plugin for agents
│   ├── genome-analizer/      # Plugin for genome analysis
│   ├── ssh-remote-connection/# Plugin for SSH
│   ├── yandex-wordstat/      # Plugin for Wordstat API
│   ├── codex-review/         # Plugin for cross-agent review
│   ├── fal-ai-image/         # Plugin for image generation
│   ├── yandex-search-api/    # Plugin for Yandex Search API
│   ├── yandex-direct/        # Yandex Direct statistics and management
│   ├── yandex-metrika/       # Plugin for Yandex Metrika analytics
│   ├── yandex-webmaster/     # Plugin for Yandex Webmaster API
│   ├── zoomkit/               # Plugin for ZoomKit API
│   ├── telegraph-publisher/  # Plugin for Telegraph publishing
│   ├── crawl4ai-seo/         # Plugin for SEO crawling
│   ├── telegram-channel-parser/ # Plugin for Telegram channel parsing
│   ├── x-research/              # Plugin for X/Twitter research
│   ├── github-pages-publisher/  # Plugin for GitHub Pages publishing
│   ├── sourcecraft-publisher/   # Plugin for SourceCraft Sites publishing
│   ├── reddit-skill/            # Plugin for Reddit API
│   ├── knowledge-compiler/      # Plugin for compiling sources into skills
│   └── perplexity-search/       # Plugin for search and research via Perplexity API
└── README.md