Pascal Playwright Mcp

Browser automation via Playwright MCP server. Navigate websites, click elements, fill forms, extract data, take screenshots, and perform full browser automat...

This skill appears to be what it claims — a Playwright MCP helper — but a few practical safety checks are recommended before you run it: 1) Install @playwright/mcp from the official npm registry and pin a specific version. 2) Run the MCP server in a restricted environment (container or sandboxed VM) if you will point it at untrusted sites. 3) Configure --allowed-hosts and --blocked-origins to prevent navigation to arbitrary domains, and avoid using --no-sandbox unless you understand the risks. 4) Be cautious with browser_evaluate (arbitrary JS): it can read page data and interact with authenticated sessions; do not evaluate untrusted scripts or send outputs to untrusted endpoints. 5) Verify the runtime configuration actually restricts filesystem access and service capabilities as the SKILL.md claims. If you need higher assurance, ask the publisher for the exact MCP server launch flags and the source/package checksum before installing.

Type: OpenClaw Skill Name: pascal-playwright-mcp Version: 1.0.0 The skill bundle provides browser automation capabilities via Playwright MCP, a legitimate and widely used tool. The `SKILL.md` clearly outlines the skill's purpose, installation steps using `npm` and `npx` for the official `@playwright/mcp` package, and explicitly mentions security features like restricted file system access, host validation, and sandboxing. The `examples.py` script is purely illustrative, demonstrating conceptual MCP tool calls by printing them, without actually executing any commands or interacting with external systems. There is no evidence of prompt injection, data exfiltration, persistence mechanisms, or other malicious intent.