[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-80676":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":15,"stars7d":16,"stars30d":16,"stars90d":14,"forks30d":14,"starsTrendScore":17,"compositeScore":18,"rankGlobal":9,"rankLanguage":9,"license":19,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":22,"hasPages":20,"topics":23,"createdAt":9,"pushedAt":9,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":14,"starSnapshotCount":14,"syncStatus":15,"lastSyncTime":27,"discoverSource":28},80676,"OpenTor","vichhka-git\u002FOpenTor","vichhka-git","Tor\u002FDark Web access skill for AI agents (OpenCode & Claude Code). Search 12 dark web engines, spider .onion sites, extract IOCs. Orchestrator-conductor architecture — zero external LLM deps.",null,"Python",50,7,47,0,2,3,6,2.71,"Other",false,"main",true,[],"2026-06-12 02:04:05","# OpenTor 🧅\n\n[![skills.sh](https:\u002F\u002Fskills.sh\u002Fbadge\u002Fvichhka-git\u002FOpenTor)](https:\u002F\u002Fskills.sh\u002Fvichhka-git\u002FOpenTor)\n\n**Tor \u002F Dark Web Access for AI Agents — OpenCode & Claude Code Skill**\n\nOpenTor gives LLMs full access to the Tor network and .onion hidden services.\nNot a standalone tool — an **orchestrator-conductor architecture** where the LLM\nis the intelligence and the Python modules provide mechanical transport, search,\nand entity extraction.\n\n```bash\nnpx skills add vichhka-git\u002FOpenTor    # install via skills.sh\n```\n\n```bash\n# Quick start — the LLM runs these:\npip install -r requirements.txt            # install dependencies\npython3 scripts\u002Fsetup.py                   # interactive setup\npython3 scripts\u002Fopentor.py check           # verify Tor\npython3 scripts\u002Fopentor.py search \"ransomware leak\"  # search dark web\n```\n\n## Install as a Skill\n\n### Claude Code\n\n```bash\n# Clone into Claude Code skills directory\ngit clone https:\u002F\u002Fgithub.com\u002Fopentor\u002Fopentor ~\u002F.claude\u002Fskills\u002Fopen-tor\n\n# Or copy from local\ncp -r OpenTor ~\u002F.claude\u002Fskills\u002Fopen-tor\n\n# Run setup (the LLM can do this itself)\ncd ~\u002F.claude\u002Fskills\u002Fopen-tor\npip install -r requirements.txt\npython3 scripts\u002Fsetup.py\n```\n\nAfter install, start a **new Claude Code session**. The skill loads automatically\nwhen you ask about dark web topics, .onion URLs, ransomware groups, credential\nleaks, or Tor-based OSINT.\n\n### OpenCode\n\n```bash\n# Copy to OpenCode skills directory\ncp -r OpenTor ~\u002F.config\u002Fopencode\u002Fskills\u002Fopen-tor\n```\n\n### LLM Self-Install\n\nThe orchestrator (Claude) can install itself. Just say:\n\n> \"Install OpenTor and set up Tor access\"\n\nThe LLM will:\n1. Clone the repo to the skills directory\n2. Run `pip install -r requirements.txt` (asking for sudo\u002Fvenv choice)\n3. Run `python3 scripts\u002Fsetup.py` for interactive configuration\n4. Install and start Tor if not present\n5. Verify with `python3 scripts\u002Fopentor.py check`\n\n### Standalone (without skill system)\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fopentor\u002Fopentor\ncd opentor\npip install -r requirements.txt\npython3 scripts\u002Fsetup.py\npython3 scripts\u002Fopentor.py check\n```\n\n## Requirements\n\n| Dependency | Version | Purpose |\n|-----------|---------|---------|\n| Python | 3.10+ | Runtime |\n| Tor | any | SOCKS5 proxy (:9050) + ControlPort (:9051) |\n| `requests[socks]` | >=2.28 | HTTP through Tor SOCKS5 |\n| `beautifulsoup4` | >=4.11 | HTML parsing for search + fetch |\n| `python-dotenv` | >=1.0 | .env configuration |\n| `stem` | >=1.8 | Tor ControlPort (circuit rotation) |\n\nNo LLM API keys required. No external AI service dependencies. The orchestrator\nIS the LLM.\n\n## Commands\n\n| Command | What it does |\n|---------|-------------|\n| `opentor.py check` | Verify Tor is running, show exit IP |\n| `opentor.py engines` | Ping 12 search engines, show latency\u002Freliability |\n| `opentor.py search \"query\"` | Search dark web — all engines, scored results |\n| `opentor.py fetch \"url\"` | Fetch any .onion or clearnet URL through Tor |\n| `opentor.py renew` | Rotate Tor circuit (new identity) |\n| `opentor.py entities --text \"...\"` | Extract IOCs (emails, crypto, onions, PGP) |\n| `opentor.py crawl \"url\"` | Spider a .onion site — follow links, map structure |\n| `opentor.py crawl-export \u003Cid>` | Export crawl results |\n\n### Options\n\n```\n--mode MODE        threat_intel | ransomware | personal_identity | corporate\n--engines NAME     Specific engines (e.g. Ahmia Tor66)\n--max N            Max results (default 20)\n--format FMT       json (default) | csv | stix | misp | text\n--out FILE         Write output to file\n--json             Machine-readable JSON output\n--depth N          Crawl depth (default 3, for crawl subcommand)\n--pages N          Max pages (default 100, for crawl subcommand)\n--stay             Stay on same .onion domain (for crawl)\n```\n\n## Features\n\n### Dark Web Search\n12 verified-live engines queried in parallel through Tor. Results scored by\nBM25 relevance, deduplicated across engines, with 30-minute SQLite cache.\n\n### .onion Spider\nBFS crawler follows links through .onion sites, extracts entities (emails,\ncrypto addresses, PGP keys, onion links), builds a link graph, stores\neverything in SQLite. The LLM cannot navigate hundreds of Tor URLs — the\nspider can.\n\n### Entity Extraction\nRegex-based IOC extraction: emails, BTC\u002FXMR\u002FETH addresses, .onion URLs,\nPGP keys, phone numbers, IPs, domains.\n\n### Output Formats\nExport results to JSON, CSV, STIX 2.1 Bundle, MISP Event — feed directly\ninto threat intelligence platforms.\n\n### Analysis Modes\nFour modes with engine routing: `threat_intel`, `ransomware`, `personal_identity`, `corporate`.\n\n### SQLite Persistence\nSearch results cached across sessions. Engine reliability tracked with\nexponential time-decay scoring. Crawl data stored for export.\n\n### Content Safety\nAutomatic blacklist for CSAM and illegal content. Cannot be disabled.\n\n### Clearnet-First Strategy\nThe skill teaches the LLM to search public internet first (Google\u002FDuckDuckGo)\nto understand context before targeting dark web queries — validated in real\nransomware investigations.\n\n### Professional Reporting Standards\nSix built-in thinking directives teach the LLM to:\n- Label every assertion (`✓ Observed` \u002F `⚡ Inferred` \u002F `❓ Uncertain` \u002F `🤖 AI Analysis`)\n- Treat folder names as hypotheses (crawl before reporting)\n- Report raw data before interpretation\n- Never fill gaps with assumptions\n- Make every finding traceable to source evidence\n\n## Roadmap\n\n- **Domain allow\u002Fdeny lists** — scope control for spider (allowlist specific .onion domains, block others)\n- **Export redaction** — strip PII, credentials, and sensitive data from STIX\u002FMISP exports\n- **Safe mode** — read-only default (search + fetch only, crawl disabled unless explicitly allowed)\n- **Crawl scheduling** — time-boxed spider runs (stop after N minutes regardless of depth\u002Fpages)\n- **Report templates** — customizable output structure per investigation type\n\n## License\n\nMIT License — see [LICENSE](LICENSE).\n\nEngine catalogue adapted from [Robin](https:\u002F\u002Fgithub.com\u002Fapurvsinghgautam\u002Frobin) (MIT).\n\n**Use responsibly.** Built for OSINT, threat intelligence, and security research.\n","OpenTor 是一个为AI代理提供Tor网络和暗网访问能力的项目。它支持搜索12个暗网搜索引擎，爬取.onion网站，并提取IOC（入侵指标）。该项目采用编排器-指挥架构，不依赖外部LLM，确保了数据处理的安全性和独立性。其核心功能包括通过Tor网络进行安全的HTTP请求、HTML解析以及自动配置Tor服务。适用于需要在保证隐私的情况下进行暗网研究或情报收集的场景，如网络安全分析、威胁情报获取等。","2026-06-11 04:01:37","CREATED_QUERY"]