Tor Browser Automation

Headless browser automation with Tor SOCKS5 proxy support for accessing .onion sites and anonymous browsing. Use when navigating dark web resources, scraping...

This skill appears to do what it says: automated browsing through a local Tor SOCKS5 proxy using Playwright. Before installing, ensure you: (1) understand and comply with laws and terms of service for sites you access (dark‑web scraping can carry legal risk); (2) run Tor in an isolated environment or VM and avoid running the tool as root; (3) verify Playwright and Tor are installed from official sources; (4) be aware the tool will run arbitrary page interactions you issue (don’t paste secrets into pages you automate); and (5) review the full script in your environment if you need higher assurance (the code is straightforward but I only reviewed the visible portions). If you require the agent to run this autonomously, consider restricting when and what it may browse to limit accidental misuse.

Type: OpenClaw Skill Name: tor-browser Version: 1.1.0 The OpenClaw AgentSkills skill bundle for 'tor-browser' is benign. It provides headless browser automation with Tor SOCKS5 proxy support, which aligns with its stated purpose of navigating .onion sites and anonymous browsing. The `SKILL.md` and `references/setup-guide.md` files contain standard setup instructions and do not exhibit prompt injection attempts. The `scripts/tor-browser.py` code uses Playwright for browser control, correctly configures the Tor proxy, and performs typical browser actions without attempting data exfiltration, persistence, or arbitrary command execution. The `check_tor_connection` function legitimately queries `check.torproject.org` to verify Tor connectivity. While powerful, the skill's capabilities are consistent with its described functionality and lack any indicators of intentional malice.