Research across dozens of companies building agents in production reveals a striking pattern: the agent harness — the thing that actually executes the agent loop — is completely commoditized. With few exceptions, everyone is running Claude Code, Open Code, or some fork of an open-source equivalent. Companies that started with custom harnesses migrated through homegrown loops to Goose to Aider to Claude Code before settling on whatever the current best option is.
There is nothing to build here. There is nothing to differentiate on. Even the companies that bet heavily on owning the harness are in an uncomfortable position as the open-source and lab-provided options improve every month. The harness is something you have as a force multiplier for an existing product. It is not something you put out to market as a standalone offering.
The real opportunity lives in the layers almost everyone is still building themselves. Look at any honest survey of how companies assemble their agent stacks and the pattern is overwhelming. The harness is bought. The LLM gateway is bought. Triggers and services are bought. But context management, memory, skills, orchestration, and session state are almost universally hand-built. The build-versus-buy chart is mostly blue, with very little green.
The existing frameworks at those layers — LangChain, Mastra, the various agent SDKs — are overcomplicated. They started full-featured instead of starting simple. What the market needs is not LangChain with more features. It is something built bottom-up: a few clean abstractions one layer upstream of the complexity, not buried beneath it.
The right question for any platform company is: what does that engineer Google when their duct-taped agent breaks at 2 a.m.? That search query is the entry point. That is the wedge. I've argued elsewhere that the market is an arena of arenas — and in that fight, the only durable position is to own the abstractions above the harness, not the harness itself.
Sources
Related Essays
The Framework Trap
LangChain pivoted to LangSmith. E2B sells the sandbox. The agent harness is not the product — it is the thing you give away. Monetization lives in the infrastructure underneath.
The Agent Harness Problem
Enterprise agents need layered interfaces, real software skills, and flexible platforms. The harness around the model matters more than the model.
Agents Are Software, and Software Needs a Factory
People talk about agent harnesses as if the harness is the interesting part. It is not. The interesting part is the factory — sandboxing, orchestration, persistence, model translation.
Key takeaways
- The agent harness is commoditized — every serious team is running Claude Code, Open Code, or some open-source fork.
- Companies betting on owning the harness as a standalone product are in an uncomfortable position. The harness is a force multiplier, not a market.
- The real opportunity lives in context, memory, skills, orchestration, and session state — layers almost universally hand-built today.
- The market does not need a more featureful LangChain. It needs clean abstractions one layer above the duct tape, not buried beneath it.
FAQ
What is the agent harness, and why is it commoditized?
The harness is the execution loop — the thing that runs the agent's tool-call cycle. Claude Code, Open Code, and their open-source forks have converged on roughly the same shape, and most teams have churned through multiple harnesses before settling on whatever the current best option is. There is no durable differentiation left to build there.
What layers above the harness are still hand-built?
Context management, memory, skills, orchestration, and session state are almost universally custom. The honest build-versus-buy chart for agent stacks is mostly blue with very little green above the harness. That blue is the platform opportunity.