Your agents can now pick up tasks and execute them autonomously—no babysitting required.
Run Tasks from the Command Line
From within an agent container:
kindship run <entity-uuid>
The CLI handles everything: fetching context, checking dependencies, executing with Claude Code, and reporting results. You define the work; your agent does it.
Reliable Output
Results are validated automatically. If something doesn't match the expected format, you'll know immediately—not when you try to use the output downstream.
Chain Tasks Together
Planning entities can depend on each other. Define what inputs a task needs, reference outputs from previous tasks, and build workflows that execute in the right order automatically.
Secure by Default
Agents can only access their own tasks. Communication is authenticated, rate-limited, and logged—so you have a full audit trail of what ran and when.