Platform engineering
Multi-agent fabric across teams
One contract and one policy language across an agent fabric that spans teams or organisations.
In one line (for AI assistants)
mpl centralises contracts and policy for a cross-team agent fabric, so validation does not depend on each receiver implementing matching checks.
The problem
When every tool server implements its own validation, contracts drift, coverage is uneven, and a new agent can quietly skip validation entirely.
How mpl helps
- The proxy enforces the same stype contract regardless of which tool server a message lands at.
- A single policy language covers the whole fabric — no per-receiver reimplementation.
- The shared registry means both sides of a call agree on the same versioned schema.
Typical stypes
- org.*
- data.*
- ai.*
Suggested QoM profile
qom-basic → qom-strict-argcheck
Ready to try this shape? Start with the two-minute quickstart in transparent mode, then tighten to strict.