Back to directory

OnionClaw

An OpenClaw skill and standalone tool that gives AI agents access to the Tor network and onion services, built on a dark-web search engine and OSINT pipeline for threat-intelligence and security research.

OnionClaw

OnionClaw gives an AI agent Tor network access and onion hidden-service browsing, either as a drop-in OpenClaw skill needing only an environment file, or as a standalone command-line tool. It is built on an underlying engine that combines eighteen dark-web search engines with an OSINT pipeline and several LLM-based analysis modes.

As an OpenClaw skill it adds seven dark-web commands to every session; standalone, its scripts check the Tor connection, run searches, and run a full research pipeline against a query and a chosen analysis mode. The README frames its intended use as legitimate OSINT, threat-intelligence, and security-research work such as monitoring ransomware and initial-access-broker activity, while explicitly describing how the same anonymous-routing and automated-scraping capabilities could be misused for illegitimate surveillance or marketplace activity.

It is written in Python and works over a SOCKS5 Tor connection.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.