Jev in production › Agent guardrails and approvals
Foreman, an agent supervisor and software factory foreman, is built directly on Jev to route tasks and check agent work (author).
Foreman watches the software factory floor with TypeSafe AI's Jev, placing a fast decision model above slower coding agents.
Give it a ticket, specification, bug report, or any free-form software job. Foreman can run Codex, OpenCode, Hermes, or LangChain's Deep Agents Code (dcode) as workers. It can also attach supervision to supported coding-agent sessions, including Pi and Pi Durable.
The coding agent does the software engineering. Foreman independently assesses completion and requirements against the original job. It also checks whether test evidence is sufficient and whether verification or human input is needed. You can configure additional checks and responsibilities for your workflow.
For the project's own README, linking back here: