I’m the founder of PlanSpec, a coordination layer for humans and AI agents.
As AI agents get better at execution, we’ve found that human coordination is becoming the bottleneck: keeping goals aligned, managing context across sessions, and knowing when work is actually “done.”
PlanSpec is an open-source, goal-oriented declarative specification for coordinating humans and agents. It lets you describe goals, plans, dependencies, and explicit gates for review or approval, so execution can move quickly without losing shared understanding.
It’s early, but usable today. We’re especially interested in feedback from teams already working with multiple agents or long-running agent workflows.
Repo + docs: https://planspec.io