Integrate and interoperate
ARPA is designed to compose with adjacent protocols and governance architectures without collapsing their trust semantics into ARPA or importing theirs as implicit authority.
Integration paths
| Integration | Guide/profile | Boundary to preserve |
|---|---|---|
| TGA agentic governance architecture | TGA Agentic Governance Alignment | ARPA is an optional control-plane realization, not the complete agentic runtime |
| A2A Agent Cards | A2A Registry Integration Guide | Publication/discovery is not authority |
| A2A protocol profile | ARPA A2A v1.0 Profile | Task completion is not ARPA authorization |
| TRQP | ARPA–TRQP architecture | TRQP remains a read-only projection interface |
| General registry exchange | Interoperability guide | Recognition and authority remain scoped |
TGA agentic-governance relationship
Trust Graph Artifacts’ Agentic Systems Architecture and Governance guide provides a protocol-neutral method for bounded jobs, authority, delegation, capability separation, governed execution, evidence, revocation and assurance. ARPA can implement selected discovery, authority-control, lifecycle, federation and evidence responsibilities within that architecture.
The relationship is intentionally non-dependent:
TSMM -> canonical semantics
TGA -> generic governance / assurance pattern
TIS -> portable executable contracts
ARPA -> optional registry / authority-control realization
The machine-readable declaration is interop/tga-agentic-governance-alignment.json and is release-gated by scripts/validate_tga_alignment.py.
TypeScript integration surface
The v0.9.5 development track provides:
- a thin Node.js HTTP service;
- an
ArpaClientfor registry discovery, resolution and authority evaluation; - A2A publication projection helpers;
- conservative Agent Card compatibility classification;
- Python↔TypeScript HTTP interoperability evidence.
See TypeScript implementation for supported and deferred behaviour.