Handbook

MCP example — Playwright snapshot (Python client)

This folder contains a small offline-oriented script that uses the LCDL MCP client stack to talk to a Playwright MCP server when you choose to run it with live tooling.

Updated

Files

File Purpose
playwright_fetch_snapshot.py Minimal fetch + snapshot path (see docstring for flags).
playwright.local.json Example transport config stub (adapt for your machine).

Policy

  • No credentials in repo files; copy local JSON to a gitignored path if it contains secrets.
  • Live Playwright MCP requires Node/npx and your own authorization context—not the default learning path.

Docs

Smoke check (optional)

pip install -e ".[mcp,dev]"
pytest -q tests -k mcp --maxfail=1

(This may skip or pass depending on optional dependencies and markers.)