Orchestrates a multi-agent consensus workflow for robust, architecturally sound software planning and risk mitigation.
Ralplan is an advanced planning skill for Claude Code that implements a structured deliberation process between Planner, Architect, and Critic agents. By enforcing a consensus-based approach (RALPLAN-DR), it ensures that technical decisions are thoroughly vetted, risks are identified via pre-mortems, and alternative solutions are considered before any code is written. It acts as a sophisticated pre-execution gate, automatically redirecting vague requests into detailed project requirements and Architecture Decision Records (ADRs) to prevent wasted compute cycles and ensure high-quality implementations through parallel or sequential execution modes.
主要功能
010 GitHub stars
02Multi-agent consensus loop involving Planner, Architect, and Critic roles
03High-risk Deliberate Mode with automated pre-mortems and expanded test planning
04Automatic generation of detailed ADRs (Architecture Decision Records) upon approval
05Intelligent pre-execution gate that intercepts and scopes underspecified requests
06RALPLAN-DR structured deliberation framework for transparent decision making
使用场景
01Scoping vague project requests into actionable, testable technical plans with clear acceptance criteria
02Designing complex system migrations or high-risk security features requiring deep architectural review
03Enforcing architectural consistency and quality standards across automated code implementation tasks