vostride/agent-qa

View on GitHub ↗

Open-source self-improving QA agent for software teams. A test harness with memory. Write tests in natural language for web and mobile. agent-qa learns from every run, adapts to UI changes, and catches regressions before you ship.

899 ★13 forksTypeScriptUpdated 1mo ago

What you need to know

Self-improving agentic QA harness with memory - write tests in natural language for web and mobile, with self-healing execution and MCP plus skills for coding agents

Install

npm install -D agent-qa
npm install -D @vostride/agent-qa-subscription-auth
npx agent-qa init and npx agent-qa install-browsers --chromium

Usage

  • Run tests from CLI with npx agent-qa run tests/example.yaml
  • Start the dashboard with npx agent-qa dashboard --open

Key features

  • Write actions and assertions in natural language for web and mobile
  • Self-healing test execution that recovers from UI drift
  • Execution memory that self-improves across runs
  • Dashboard and CLI for humans plus MCP and skills for agents
  • Action cache reusing validated plans to cut tokens
  • Sandboxed hooks in Docker for Node, Bun, Python and Bash
  • Bring your own LLM via OpenAI/Anthropic-compatible endpoints

Best for

QA teams who want LLM-driven, self-healing tests with cross-run memory

Caveats

  • Docker required when using hooks
  • Mobile projects need driver installation
Platforms: LocalClients: Codex · Claude Code

Reviewed 2026-08-11

Topics

ai-agentsai-testinganthropicautonomous-agentsbrowser-automationchatgptclaude-codeclawdbotcodexdeveloper-toolsdsh-pluginend-to-end-testingfrontendmcpmcp-servermcp-toolsmodel-context-protocolopenaiplaywrightqa-automation
Stars
899★
Forks
13
Language
TypeScript
License
NOASSERTION
Created
2026-05-07
Last push
2026-08-03