Standalone Go runner and @9l/playwright SDK for bounded Playwright execution with validated receipts.
Run Playwright tests with bounded execution and attributable receipts. Add named steps and natural-language goals while keeping explicit assertions as the test oracle. Local execution needs no QualityMax account or hosted service.
9lives runs, assesses, and stress-tests Playwright tests locally; qmax-code is the terminal coding and QA agent that drives broader repository workflows.
Documentation · Explainer · Local demo · npm SDK · CLI releases
A deliberately stale button name breaks a Playwright test against the public QualityMax Demo Bank. Follow the draft through assessment, failure, offline repair and a passing candidate.
The GIF is a 68.5-second edited replay of real browser captures and checked CLI results, with large text and one step per frame. Experimental healing changed one locator literal, preserved every assertion and verified one passing test. The candidate was saved for review; the original file was not changed.
Run this example · Still image · Checked results
- Agent-assisted test generation: your coding agent creates Playwright specs using the optional QA skill. Record creation provenance, assess the draft and execute it; there is no native 9l generate command. - Assessment: 9l assess reviews waits, skips, missing or unawaited assertions, and requirement mappings. Findings are advisory. - Execution: plan, run, status, result and cancel bound execution and retain attributable terminal receipts. - Healing: experimental native 9l heal preserves assertions and verifies supported locator-only candidates in isolation before acceptance. - Fault testing and c
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/Quality-Max--9lives-runner.html)
Model your API as a graph once. Long-chain integration tests, layer × environment matrices, CI-ready runs, and an MCP server — all from the same YAML.
magna-nz/tallybookPrices your Claude Code and Codex sessions from the transcripts on disk, shows what each agent cost over any period, and tells you in plain English what would have been cheaper. CLI, Browser UI and MC
Smitner-Studio/facetRender agent-written Markdown as a live local page, and send the human's click back to the agent as typed data. Human-in-the-loop approval gates for coding agents.
vtimame/istok.shLocal-first project state, task tracking, and context for coding agents.
nassim-arifette/jevgrepJev-powered semantic code search for coding agents — find behavior across repositories via CLI or MCP, with exact source excerpts and line numbers.
jimoto-no-llm/rustdshFast Rust launcher for DeepSeek Harness, with clear Claude Code/Codex migration, model setup and lightweight local tools
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.