ralplan

What problem does it solve? Vague or underspecified requests sent directly to execution agents waste cycles on scope discovery and produce misaligned work. Ralplan intercepts these requests and routes them through a structured consensus planning workflow so execution starts with a clear PRD, testable acceptance criteria, and an approved architecture. ## Core Features & Use Cases - Consensus Planning Loop: Iterates Planner, Architect, and Critic subagents (up to 5 rounds) until the Critic approves, producing a durable consensus artifact with ADR, risks, and verification steps. - RALPLAN-DR Deliberation: Generates a structured summary of principles, decision drivers, and viable options, with a deliberate mode adding pre-mortem scenarios and expanded test plans for high-risk work like auth, migrations, or compliance changes. - Pre-Execution Gate: Detects underspecified execution prompts (e.g., "ralph add authentication") and redirects them into planning, while letting concrete prompts with file paths, issue numbers, or acceptance criteria pass through. - Use Case: A user types "autopilot build the app"; the gate redirects to ralplan, which drafts a PRD and test spec, runs Architect and Critic reviews, and hands off the approved plan to $ultragoal or $team for execution. ## Quick Start Ask the assistant to run $ralplan with a task description, optionally adding --interactive for approval prompts or --deliberate for high-risk work.