Agent Harness Taxonomy — Six Architectures for AI Agents
The agent harness determines the behavior envelope, not the model. Six agent harness architectures — and why the sixth breaks what the others assume.
Agent harness design, context engineering, agentic infrastructure, and the emerging shape of agentic organizations. Practitioner-positioned, research-grounded.
The agent harness determines the behavior envelope, not the model. Six agent harness architectures — and why the sixth breaks what the others assume.
Anthropic named a memory consolidation feature 'Auto Dream.' The metaphor wasn't decoration — it was a spec. Wittgenstein explains why.
Anthropic's Claude Code source leaked — 512K lines, every agent pattern exposed. The model layer fell to distillation. The harness layer fell next.
The agentic operating manual is the missing governance layer between your org and its AI agents. Four layers that replace coordination overhead.
When AI makes capability free, trust becomes the only scarce input. That reshapes organizational theory from Coase forward.
Agent Harness Taxonomy — Six Architectures for AI Agents
2026-04-07•12 min read•analysis
The Trust Contraction — When Trust Is the Only Scarce Input
2026-04-07•7 min read•essay
The Agent Harness Layer Is Indefensible — The Claude Code Leak
2026-04-01•10 min read•analysis
Permissionless Intelligence — Proof of Work, Proof of Trust
2026-03-25•7 min read•essay