Implementation Agent (Code) v4.2.4

Automate the software development lifecycle using a 7-phase AI flow: Research, Plan, Test, Code, Debug, Verify, and Document. Provide a high-level goal, and the agent manages state via RunContext to generate, test, and document code consistently.

Test-driven development is supported through diff-driven testing strategies within the 7-phase flow. The agent writes tests before coding, executes verification, and employs self-healing mechanisms to classify failures and automatically retry non-deterministic errors.

Yes, the code generation agent classifies build and test failures using self-healing mechanisms. It automatically retries failures and applies circuit breakers for deterministic errors, logging all outcomes for KI Learning to improve future error recovery.

Maintain shared state across testing, debugging, and code generation phases using RunContext integration. It ensures consistency and traceability throughout the entire 7-phase development flow, linking research, planning, and documentation outputs seamlessly.

No external dependencies are required to run this software development agent. It operates independently using bundled scripts, references, and assets, executing the full development lifecycle directly within your environment.

Use a structured 7-phase development flow over direct AI code generation to ensure comprehensive research, planning, and verification. This approach integrates debugging and documentation, delivering higher quality code with automated error recovery and outcome logging.