ralplan

To plan complex software architecture changes, you need a structured workflow that generates independent Planner, Architect, and Critic reviews. This process produces consensus-approved specifications, dependency graphs, and risk registers before coding begins.

Preventing flawed implementation work requires an adversarial planning process with technical pushback and explicit approval gates. Replacing simulated consensus with independently produced reviews ensures potential issues are identified before execution starts.

A multi-agent consensus review workflow functions by coordinating isolated Planner, Architect, and Critic agents through a strict iterative state machine. It generates filesystem checkpoints and requires testable acceptance criteria to approve migration plans.

Yes, you can use this planning process for production incidents and public API changes. It applies adversarial review, dependency analysis, and risk assessment to complex software engineering tasks to ensure safe execution.

Brainstorming requirements is necessary before creating an implementation plan. The workflow surfaces open questions and records answers to ensure all ambiguities are resolved before generating the final consensus-approved plan.

You should not use a multi-agent consensus planning approach for simple tasks that do not require complex dependency analysis or risk assessment. It is designed for authentication changes, migrations, and public API modifications needing adversarial review.