Date: August 17, 2025 Status: Superseded — never executed (status corrected 2026-08-29) Deciders: Principal Architect, Chief Architect, Chief of Staff Classification: Strategic (Long-term Vision)
MCP has proven successful as our integration layer, but it’s a general-purpose protocol not optimized for Product Management workflows. Through implementing PM-033c/d, we’ve discovered patterns and requirements unique to PM orchestration that aren’t adequately expressed in standard MCP.
The phrase “downpayment on our own protocol” emerged during architecture discussions when we realized we’re not just consuming protocols—we’re learning what a PM-specific protocol needs to enable.
As noted in our strategic sessions: “Good architecture naturally becomes open because clean abstractions are shareable, well-defined boundaries become protocols.” Our architecture has reached the maturity where defining our own protocol extensions is both possible and necessary.
We will invest in developing PM-specific protocol extensions and, eventually, a standalone PM orchestration protocol, building on MCP’s foundation while addressing Product Management’s unique requirements.
Phase 1: MCP Extensions (Current - 3 months)
Phase 2: Protocol Specification (3-6 months)
Phase 3: Ecosystem Development (6-12 months)
What MCP Lacks for PM:
piper_protocol:
extends: mcp
version: 0.1.0
primitives:
- workflow_state
- approval_chain
- velocity_metric
- dependency_graph
- resource_allocation
- risk_assessment
patterns:
- sprint_transition
- story_decomposition
- stakeholder_notification
- metric_aggregation
- conflict_resolution
guarantees:
- eventual_consistency
- audit_trail
- rollback_capability
- privacy_preservation
Why Rejected: MCP is too generic. PM needs would always be second-class, requiring workarounds and custom implementations that fragment the ecosystem.
Why Rejected: Closed protocols limit adoption. Our architecture philosophy (“good architecture becomes open”) demands openness. Network effects require community participation.
Why Rejected: No PM protocol leader has emerged. Waiting cedes competitive advantage. Our implementation experience positions us to lead.
piper-protocol GitHub repositoryThe investment in protocol development represents a strategic bet that PM tooling will converge on standards, and that the first credible standard will capture significant value. Our unique position—having built deep integrations across multiple PM tools—gives us the experience to define what this protocol needs.
The “downpayment” isn’t just code or documentation—it’s the commitment to open, shared infrastructure for Product Management. Every pattern we discover, every integration we build, contributes to the protocol specification.
As our architecture discussions revealed: “Bad architecture can never become open because it can’t bear the scrutiny.” The fact that we’re ready to open our patterns as protocol proves our architecture’s quality.
Status corrected Accepted → Superseded (never executed). No phase of the Piper Protocol plan (spec, reference implementation, open-sourcing) was ever built; no decision record ever cancelled it — this note is that record. The distribution question this ADR reached for is answered by PDR-005/PDR-006 (BYOC + hosted MCP) using the industry’s protocol (MCP) rather than an owned one.