QA Radar
Prioritize which files and tests an agent should run first using churn, coverage gaps, and test-mapping risk scores.
Overview
QA Radar is an MCP server for the Ship phase that ranks which files to test first using churn, coverage, and test-mapping risk scores.
What is this MCP server?
- Risk scores blend git churn, coverage holes, and existing test mapping
- Steers Claude/Cursor agents toward high-impact files instead of random test runs
- PyPI package qaradar 0.3.5 with stdio MCP transport
- Suited for repos where test suites are large and timeboxed
- MCP integration surface, not a standalone test runner
- Server version 0.3.5 on PyPI identifier qaradar
- Repository: github.com/Muratkus/qaradar
What problem does it solve?
Agents and humans waste ship cycles running broad test suites without knowing which recently changed, under-covered files actually matter.
Who is it for?
Indie builders with git history and coverage data who want agents to triage test work before each release.
Skip if: Greenfield projects with no tests, no coverage tooling, or teams that only need a single fixed smoke suite.
What do I get? / Deliverables
Your agent gets ordered, risk-scored file targets so pre-release testing focuses on the highest-impact gaps first.
- Risk-ranked list of files to test first
- Agent-actionable churn, coverage, and test-mapping context
- Focused pre-ship test plan instead of blanket suite runs
Recommended MCP Servers
Journey fit
QA Radar sits in Ship because its core job is test prioritization and risk scoring before release, not greenfield feature work. Testing is the canonical shelf: churn-to-coverage mapping directly answers what to test first in a pre-ship QA pass.
How it compares
Test-prioritization MCP server, not a full CI runner or coverage dashboard UI.
Common Questions / FAQ
Who is QA Radar for?
Solo and small-team developers using AI coding agents who already have tests and want smarter prioritization before ship.
When should I use QA Radar?
Use it during Ship/testing when churn is high, coverage is uneven, and you need the agent to target the riskiest files first.
How do I add QA Radar to my agent?
Install the qaradar PyPI package, register the stdio MCP server in Claude Code or Cursor per your client’s MCP config, and point it at your repo.