Facilitate cross-model debates between AI models for design and architecture validation.
npx skills add https://github.com/ggombee/code-forge --skill debate-ggombee
Please help me install this Agent Skill. Skill: debate Source: https://github.com/ggombee/code-forge/tree/main/skills/debate Command: npx skills add https://github.com/ggombee/code-forge --skill debate-ggombee
This Skill facilitates multi-angle validations through cross-model debates, enhancing design decisions and architecture selections.
Initiate a debate on a technical decision by running /debate.
High-intent search queries and answers about installing and using this skill.
Cross-model debate validates technical architecture decisions by facilitating discussions between different AI models to remove single-viewpoint blind spots. This process improves design decision justification and architecture selection robustness through multi-angle validation.
You do not need Codex CLI to run an AI architecture debate. While Codex CLI integration enables 1:1 debate with Claude, the Skill provides a self-debate option using internal Claude capabilities for technical validation without external setup.
Start a multi-model debate for design decision-making by running the /debate command. This initiates the Skill to facilitate real-time parallel debates using Agent Teams or direct Claude and Codex CLI integration.
Self-debate uses internal Claude capabilities for technical validation without extra setup, while parallel debate utilizes Agent Teams for real-time multi-angle validation. Self-debate offers immediate feedback, whereas parallel debate removes single-model blind spots.
Agent Teams can be used for real-time AI architecture validation by facilitating parallel debates. This approach applies multi-angle validation to design decisions, ensuring architecture selections are robust and justified across different AI models.
Use cross-model debate for design decisions when you need to validate architectures and remove blind spots from a single viewpoint. It is specifically designed for scenarios requiring robust technical decision-making and multi-angle justification.
zhjaiOrchestrate multi-agent debate and evidence checking among heterogeneous AI agents.
racciolyGenerate end-to-end technical implementation plans from feature specifications.
DmitriyYukhanovOrchestrate concurrent code reviews between Claude agents and Codex CLI.
valorisaCoordinate multi-LLM debates with a silent stress-test phase to validate project plans.
