Should coordination protocols wrap existing systems or replace them?

The agent-protocol landscape is already crowded: MCP for model-to-tool access, A2A for agent-to-agent task collaboration, A2UI for interface delegation, DIDComm for secure DID-based messaging, ANP for discovery and negotiation, UCP for agentic commerce. Each covers a real slice of the space. The Foundation Protocol's design choice is to not add a competing slice but to provide the shared substrate these ecosystems keep re-creating in different forms — a graph-first control plane that lets them compose across boundaries while preserving identity, authority, and accountability. By separating a small protocol core from profiles, extensions, and bridges, it enables incremental adoption: you keep your existing protocols and bridge them in, rather than migrating.

This is a strategic pattern about how infrastructure standards win. Replacement demands that an entire ecosystem abandon working investments simultaneously, which rarely happens; bridging lets value accrue at the margin as each protocol connects. The counterpoint is that a wrapping layer can become a lowest-common-denominator abstraction that loses what made each underlying protocol sharp, and a bridge adds a translation surface that can itself fail or be attacked. But for a fragmenting agent ecosystem, composability beats purity. This matters because it predicts which coordination standards survive: those that reduce integration and governance overhead without forcing a rewrite.

— "Foundation Protocol: A Coordination Layer for Agentic Society", https://arxiv.org/abs/2605.23218