Autonomous Testing Agent

What problem does it solve?

Teams often ship regressions because test coverage lags behind code changes, and writing or updating tests is slow and error-prone; this skill closes those gaps by autonomously discovering missing coverage, creating tests, executing them, and iterating until failures are understood and resolved.

Core Features & Use Cases

• Source-driven discovery of test gaps: Detects project type and identifies uncovered functions, API endpoints, web components, and low-coverage branches, then maps existing test coverage.
• AI-written test generation with model routing: Generates test scaffolds for uncovered areas using different model tiers based on complexity and quality criticality (e.g., auth/security).
• End-to-end execution, evaluation, and automated repair loop: Runs the appropriate test tooling (pytest/vitest/playwright), classifies failures (TEST_BUG, CODE_BUG, ENV_BUG), and applies fixes with bounded retries.
• Structured reporting and dashboard integration: Produces a machine-readable report including pass/fail counts, coverage, auto-fixed vs manual-needed items, and prioritized next actions.

Quick Start

Ask Maggy to run the full autonomous test cycle by executing: maggy test autonomous.