A paper posted to arXiv in December proposes folding verifiable-credential exchange directly into Google's Agent2Agent protocol rather than treating identity as a separate add-on system like LOKA or ANS. The approach uses the DIF presentation proof protocol combined with JSON-LD-encoded verifiable credentials layered on top of A2A messages themselves. The design splits the world into "security domains." Each agent controls its own decentralized identifier and its own verifiable credentials, and every agent in a given domain is deployed by a dedicated orchestrator responsible for that domain. Critically, each agent's DID is anchored not in a private ledger controlled by one vendor, but in a distributed ledger jointly operated across multiple security domains — so no single company or cloud provider holds the root of trust. This matters because most agent-identity proposals so far have imagined a new protocol layer sitting alongside A2A and MCP: LOKA's Universal Agent Identity Layer, or the various ANS-style naming directories, all assume agents will speak a separate identity dialect before or after they speak A2A. This paper instead argues the trust establishment should be native to the interoperability mechanism agents already use to hand off tasks to each other. Instead of an agent presenting its papers at a separate checkpoint, the credential exchange rides along inside the same A2A exchange that carries the task itself. The practical difference shows up at multi-domain boundaries. When an agent deployed in one organization's security domain needs to hand a task to an agent in a different organization's domain, the receiving agent can request a DIF presentation proof as part of the normal A2A handshake, verify it against the shared ledger, and decide whether to proceed — without needing a side-channel identity broker or a separate registry lookup. The orchestrator-per-domain model also means revocation and lifecycle management stay local to whoever deployed the agent, while the anchoring ledger keeps that agent's DID resolvable to outside domains. The paper is explicit that this is a design proposal building on prior work, including LOKA and authenticated-delegation schemes that combine verifiable credentials with OIDC. Its contribution is narrower and more concrete than those broader frameworks: a specific binding of existing DIF and W3C standards onto an existing, already-deployed interoperability protocol, rather than a new protocol stack. Whether A2A implementers actually adopt credential exchange as a required field, versus an optional extension nobody turns on, is the open question the paper doesn't resolve.