Instead of sending accessibility tree snapshots on every action, Claude just writes Playwright code and runs it. You get back screenshots and console output. That's it.
314 lines of instructions vs a persistent MCP server. Full API docs only load if Claude needs them.
Same browser automation, way less overhead. Works as a Claude Code plugin or manual install.
Token limit issue: https://github.com/microsoft/playwright-mcp/issues/889
Claude Skills docs: https://docs.claude.com/en/docs/claude-code/skills
This might be sufficient for an independent contractor or student. It shouldn't be used in a production agent.
And for privacy screenshots stay local in /tmp, but console output and page content do go to Claude/Anthropic. It’s designed for dev environments with dummy data, not prod. Same deal as using Claude for any coding help.
MCPs themselves may provide access to tools that are either deterministic or not, but the LLM using them generally isn't deterministic, so when used by an LLM as part of the request-response cycle determinism, if the MCP-provided tool had it, is not in a feature of the overall system.
SKILLS.md relies on a deterministic code execution environment, but has the same issue. I'm not seeing a broad difference in kind here when used in the context of an LLM response generation cycle, and that’s really the only context where both are usable (MCP could be used for non-LLM integration, but that doesn't seem relevant.)
The agentic system that uses MCP (e.g., an LLM) is fundamentally non-deterministic. The LLM's decision of which tool to call, when to call it, and what to do with the response is stochastic.