Skip to main content

YieldOS

YieldOS is a security layer for AI coding agents: executable policies, oracle contracts, and proof-of-fix evidence before risky actions ship.

Open app
  • AI Security
  • Agent Tooling
  • Hackathon
YieldOS landing page for safe coding with AI agents

The problem

AI coding agents can move quickly through dependency installs, credentials, file edits, commits, and deployment-adjacent actions. The missing layer is not another chat prompt. It is a runtime contract that tells the agent what is allowed, what must be reviewed, and what proof is required before the action proceeds.

YieldOS policy and oracle contract launch surface

Executable security contracts

YieldOS frames security as contracts the agent can execute against: policy gates, oracle contracts, tamper-evident events, and proof-of-fix evidence. The product direction is a control plane where non-security users can understand why an action is blocked and what evidence would make it safe.

Hackathon scope

Built during Platanus Hack 26, the project focused on a crisp risk boundary: protect the workspace before an AI agent does something that is hard to audit after the fact. My work centered on product architecture, agent safety framing, and a launch surface that makes the technical contract understandable.