AI Test Agent

An AI-powered test automation agent that analyzes code, generates tests, and produces detailed reports.

Features

• Project Analysis: Analyzes project structure and identifies components for testing
• Test Generation: Generates comprehensive tests for various programming languages
• Test Execution: Runs tests and collects results
• Reporting: Produces detailed test reports in multiple formats
• Coverage Analysis: Generates code coverage reports
• CLI Interface: Command-line interface for easy integration
• Docker Support: Containerized for easy deployment

Supported Languages

• Python
• JavaScript/TypeScript
• Java

Installation

Option 1: Pip / Pipx (recommended)

Install the CLI globally so it is available from any project directory.

pipx install ai-test-agent
# or, if you are working from a checked-out repo
pipx install .

You can also install into an existing virtual environment:

pip install ai-test-agent

Option 2: Using Poetry (development)

git clone https://github.com/yourusername/ai-test-agent.git
cd ai-test-agent
poetry install

Using Docker

docker build -t ai-test-agent .
docker run -it ai-test-agent

Quickstart Workflow

1. Initialize your project
   From inside the project directory you want the agent to manage:

   ai-test-agent init
   

   This creates .aitestagent/config.json with sensible defaults (paths, thresholds, model names).
   Rerun init at any time to regenerate or adjust the manifest.

2. Analyze the codebase
   When inside an initialized project, you no longer need to repeat --project-path:

    ai-test-agent analyze
   

   Results are written to the configured analysis output path (defaults to analysis.json).

3. Generate tests

   ai-test-agent generate
   

   Tests are written to the configured tests directory (defaults to tests/).

4. Execute tests

   ai-test-agent run
   

   Test results and coverage thresholds respect the manifest configuration.

5. Produce a report

   ai-test-agent report
   

6. Run everything end-to-end

   ai-test-agent all
   

If you prefer to override any setting ad-hoc (e.g., a different output path), pass the flag and it will override the manifest for that invocation.

Commands At a Glance

From outside an initialized project, you can still supply paths explicitly, e.g.:

ai-test-agent analyze --project-path /path/to/project

For a detailed option reference, see docs/CLI_REFERENCE.md.

Project Manifest

The manifest created by init lives at .aitestagent/config.json and stores relative paths and defaults for:

• project root
• tests/analysis/results/report output paths
• coverage thresholds
• default LLM model

Feel free to edit this file directly or rerun ai-test-agent init to regenerate it.
All commands look for the manifest in the current directory or any parent directory.

Configuration

Additional runtime settings can be provided via environment variables (particularly for LLM access):

• OLLAMA_HOST: Host for Ollama service
• OLLAMA_PORT: Port for Ollama service
• DEFAULT_MODEL: Default LLM model to use

The CLI will also create .aitestagent/.env.example during init for convenience.

Contributing

• Fork the repository
• Create a feature branch
• Make your changes
• Add tests
• Submit a pull request