Deterministic agent runtime with explicit plans, typed tools, permissions, and replay. Implements Google DeepMind's Intelligent AI Delegation framework. Named after ELP's Karn Evil 9.

A1 – Know Your Agent Verifiable Authorization for Multi-Agent AI

How to lead AI agents effectively - Captain/Architect/Abdicator framework from Manus backend lead

Lightweight governance for AI delegation: decide, by risk, what an AI may settle on its own and what a human must ratify. Code review is the first worked example.

One command, three models — wire Opencode (GLM), Codex (GPT), and Antigravity (Gemini) into Claude Code as a routed delegation stack

Claude Code plugin: delegate coding tasks to Opencode as background jobs — second-model reviews, Q&A, and implementation without inflating your context window