trust-systems-meta-model

A2A Binding

This binding maps A2A-class protocol concepts into TSMM governance abstractions. It does not restate the A2A protocol. It identifies the trust-system semantics that TSMM can model, validate, and compare.

Binding posture

A2A pattern TSMM abstraction v0.19.0 treatment
Public Agent Card Service Descriptor Public descriptor disclosure
Authenticated extended Agent Card Service Descriptor + Discovery Governance Conditional disclosure, integrity, access, freshness
Agent Skill Skill Contract + Capability Negotiation Capability is negotiated before execution
Extension declaration Extension Contract + Capability Negotiation URI, version, requiredness, opt-in, failure behavior
Task Interaction Task + Task Evidence Lifecycle State transitions produce evidence obligations
auth-required state Authorization Checkpoint Authority boundary requiring explicit evidence
Message/artifact separation Content Provenance Policy + Evidence Artifact Communication is not evidence unless captured and referenced
Streaming/async/push Observability Mode Delivery mode affects replay, auditability, and user awareness
Agent opacity Opacity Boundary Hidden internals require compensating evidence

Generalized TSMM artifacts

Binding requirements

An A2A-class implementation mapped into TSMM SHOULD produce evidence for:

  1. descriptor discovery and freshness;
  2. descriptor integrity or explicit integrity waiver;
  3. authenticated extended descriptor access decision;
  4. capability negotiation outcome;
  5. required extension compatibility;
  6. authorization checkpoint for privileged task continuation;
  7. task completion artifact binding;
  8. streaming or push observability obligations where applicable.

Maturity status

This binding is candidate in v0.19.0. It is suitable for modeling, validation experiments, enterprise architecture review, and assurance design. It should not be treated as a conformance claim against the A2A protocol itself.

This site is open source. Improve this page.