This post was originally published on anton.qa. The canonical version lives there.
Playwright MCP v0.0.73 fixes a critical gap where extension channels and executable paths could not be resolved from CI/CD environment variables.
If you run Playwright MCP in Docker, Kubernetes, or ephemeral CI workers, this release removes a class of environment-specific debugging that typically consumes 15–30 minutes per incident.
What changed
Two interconnected bug fixes:
-
Extension
channelandexecutablePathnow resolve from CLI flags and environment variables (#40572) -
--browserchannel flags now propagate on--extensionpaths (#40567)
Combined, these changes mean your Playwright MCP setup can now be fully environment-driven.
The pattern
export PLAYWRIGHT_BROWSERS_CHANNEL=chromium
export PLAYWRIGHT_EXTENSION_PATH=/path/to/browser-extension
npx playwright test
The resolution hierarchy is now:
- CLI flags (highest priority)
- Environment variables
- Config file defaults
- Built-in channel defaults
MCP Registry listing
Playwright MCP is now published to the official MCP Registry on each release. This simplifies enterprise procurement and governance for teams evaluating AI-assisted testing infrastructure.
The gotcha
Environment variables set in your shell may not propagate to the MCP process spawned by your AI tool. Test this before deploying to production.
For the full breakdown — including CI/CD examples and the subprocess propagation fix — read the canonical post at anton.qa.
Anton Gulin is an AI QA Architect. Former Apple SDET, now Lead Software Engineer in Test. anton.qa
Top comments (0)