machug

Systematically verify factual claims in documents using multi-LLM triage and source-grounded verification via MCP servers and web search.

Multi-LLM adversarial spec debate. Refines PRDs and tech specs through iterative critique until all models reach consensus.

Multi-LLM adversarial project review. Evaluates progress, tech stack, code quality, CI/CD, and plan drift through parallel critique from multiple AI models.

Pi CLI wrapper — code review / adversarial challenge / consult with session continuity. Provider-agnostic second opinion via @mariozechner/pi-coding-agent.