Adopting AI coding agents is easy to start: install a tool, open a terminal, and give it a task. Making those agents part of a shared engineering workflow introduces broader challenges. Which agent should handle each task? Where should it execute? How do teammates share context, apply consistent policies, and understand usage across sessions?
In this talk, we explore those questions through Omnigent, an open-source meta-harness that provides a common orchestration layer across coding agents. We walk through the architecture and tradeoffs behind smart routing, sandboxed execution, session management, policy enforcement, and usage tracking. We also show how shared sessions and access across terminal, browser, and mobile let engineers collaborate on agent work and supervise it wherever they are.
We then bring these capabilities together through Otto, an automated issue-resolution pipeline built on Omnigent. Following an issue from reproduction through implementation, AI review, and automatic merge, we show how a common agent platform supports a workflow spanning multiple stages. We also examine health monitoring and where engineers need visibility and control as automation takes on more responsibility.
Key Takeaways:
- Design a common platform across coding agents: understand the boundaries between orchestration, individual harnesses, and execution environments.
- Make agent work manageable across a team: combine routing, sandboxes, session sharing, policies, and usage tracking to support collaboration and oversight.
- Build automation on that foundation: connect issue reproduction, implementation, review, and merge into a pipeline that engineers can inspect and monitor.
