craft

Multi-agent task execution is orchestrated by coordinating isolated agents to execute a task-graph-driven plan, enforcing strict RED/GREEN/VALIDATE gates across distinct phases. This approach ensures disciplined, test-first implementation with auditable execution logs and session artifacts.

RED/GREEN/VALIDATE gate enforcement is a strict quality control mechanism for task-graph execution that requires passing specific testing, implementation, and validation phases before proceeding. It sequences remediation and re-validation tasks automatically whenever validation fails.

To start implementing a selected plan phase, provide the required task-context per task, agent prompts, and tracker integration. The system then coordinates the isolated agents to execute the phase and produces auditable execution logs and session artifacts.

Yes, this handles remediation workflows by automatically sequencing remediation and re-validation tasks when validation fails. It coordinates isolated agents to address the failures and re-evaluate the implementation against the strict quality gates.

Task-graph-driven end-to-end implementation requires task-context per task, agent prompts, and tracker integration. These inputs enable the orchestration of isolated agents and the production of auditable execution logs and session artifacts.

Yes, this workflow explicitly supports multi-agent collaboration by orchestrating execution with clear phase boundaries and audit trails. It coordinates isolated agents across the plan to ensure disciplined, test-first execution and comprehensive session artifacts.