Warp builds self-improving agents on Claude
Warp, an AI-powered terminal and agentic development environment, has created self-improving agents using Claude's Agent Skills framework. The key insight: feedback to agents typically disappears when sessions end, so Warp built a loop where human feedback on agent output is captured and used by an 'improver' skill to update a base skill, which then improves future agent performance. This approach, now running across Warp's open-source repo, turns stateless feedback into compounding improvements, making agents more reliable and effective.
“File-based skills are a way of encoding knowledge for agents without putting that knowledge directly in the prompt, as something the agent can simply look up in the course of doing its job,” says Zach.