Piper Morgan Roadmap v16.0

Date: May 10, 2026 Author: PPM, with leadership review (PA, CXO, CIO, Architect, Lead Developer) Status: Active. Adopted May 10, 2026 (CEO ratified). M2 sub-epics M2a/b/c/d-MVP closed; M2e dispositions walked; M2f gated on canonical-retest baseline (met May 9). Supersedes: v15.0 (April 11, 2026) — archived at docs/internal/planning/historical/roadmap-v15.0-2026-04-11.md

Executive Summary

M2 sub-epics M2a/b/c/d-MVP all CLOSED; M2e gameplans walked May 3 with PM dispositions captured; M2f preceded by canonical-retest baseline gate (Run 5 May 9, 68.3% quality with v2.4-pending recalibration). #992 ETHICS-ACTIVATE arc CLOSED end-to-end (Phase A → Phase F flag-flip merged Apr 30 deecc816); ADR-061 v1.0 ratified May 4 (CEO verbal). All seven leadership roles + Lead Dev now on Code (migration arc complete Apr 26); post-migration parallel-velocity payoff demonstrated through April → May.

Methodology-corpus compounding at sub-week cadence since Apr 26: Methodology-24 (Branch-or-Anchor), Methodology-25 (Workstream Review Cadence), Pattern-063 (Parallel-Authoring Drift, Emerging), Pattern-064 (Extension Without Integration, Emerging), Pattern-065 (TBD candidate per CIO sweep), Colleague Test v2.0 → v2.4 (in flight), branched UI Lifecycle Verification Rubric v0.1, PPM Review Gates (5-class review surface, CEO-approved May 10), mailbox-discipline + sign-off discipline norms hook-enforced in CLAUDE.md. Methodology-to-runtime latency now hours-to-days, not weeks.

BYOC distribution model now in active discovery — PDR-005 scoping outline distributed May 4 to PA + Architect + CXO; eventual ADR slot TBD when Architect files (paired-document approach per PDR-001/ADR-060 precedent). HOST 360 cohort surfacing identified BYOC as “the strongest decision-debt signal in the cohort” — discovery thread responds to the surfacing.

Key changes from v15.0:

Current Position

Inchworm Position: 4.4.4 (M2 sub-epics advancing; M2d MVP CLOSED; M2e in detailed gameplan)

1. ✅ The Great Refactor (GREAT)
2. ✅ CORE functionality
3. ✅ ALPHA testing (v0.8.0 → v0.8.5.1)
4. 🎯 Complete build of MVP
   4.1. ✅ B1 - Beta Enablers (v0.8.3.x) - COMPLETE Jan 11
   4.2. ✅ A20 - Alpha Testing round 2 (v0.8.4.x) - COMPLETE Jan 18
   4.3. ✅ MUX: Modeled User Experience - COMPLETE Jan 27
   4.4. 🎯 MVP Sprints (M0 → M5) ← CURRENT
        ✅ M0: Conversational Glue — COMPLETE (v0.8.6, Mar 4)
        ✅ M1: Foundation — COMPLETE (gate closed Apr 11, 7/9 Gate 1 PASS)
        🎯 M2: Conscious Floor + Action Handlers ← MID-SPRINT
            ✅ M2a Foundation cleanup
            ✅ M2b Test infrastructure
            ✅ M2c Conversational depth (72.1% quality, post-#950 iter 2)
            ✅ M2d MUX Lifecycle MVP CLOSED end-of-day May 3
            🎯 M2e Integrations (gameplans walked May 3; #790 already shipped; #869/#900/#1039/#1040 pending)
            ○ M2f Security + Infrastructure (gated on canonical-retest baseline)
        ○ M3: Artifact Persistence (post-MVP-related issues filed during M2 walkthrough surface as M3 candidates; not yet scoped)
        ○ M4: Trust + Learning
        ○ M5: Distribution + Polish (BYOC discovery thread in flight)
5. Beta testing on 0.9
6. Launch 1.0

The Differentiator Stack (Vision V2.3 — Stable)

The MVP sprints are organized around four differentiators that, together, make Piper a colleague rather than a chatbot wrapper:

  1. Context Methodology — Five-layer model operationalized as practiced discipline
  2. Conscious Floor — LLM responses embodying grammar, Five Pillars, anti-flattening + Investment-pillar extension (#950 v0.1, May)
  3. Artifact Persistence — Conversation outputs that outlive the conversation, with composting lifecycle (M3 territory)
  4. Trust-Graduated Experience — Earned proactivity through demonstrated value (M4 territory)

Indoor plumbing (commodity, not differentiating): GitHub/Slack/Calendar/Notion via MCP plugins, file storage via SQLite + PostgreSQL audit_log persistence, auth via standard OAuth, LLM provider management via three-way fallback chain (Anthropic → Gemini → OpenAI).

MVP Sprint Status (May 10, 2026)

M0 — Conversational Glue ✅ COMPLETE (v0.8.6 Mar 4)

27 issues (5 planned + 22 discovered via Assembly Assumption — the canonical Pattern-062 instance).

M1 — MVP Foundation ✅ COMPLETE (Apr 11)

Gate 1 7/9 PASS. ADR-060 (Floor-First Routing), ADR-059 (Workflow Dispatcher), PDR-004 (Experience Philosophy). Conversation continuity (#922) folded into M2.

M2 — Conscious Floor + Action Handlers 🎯 MID-SPRINT

Theme: Make the conscious floor reliable and the action gate clean. Differentiator-stack pillars 1–2 become operational.

M2a Foundation cleanup ✅ COMPLETE (Apr 11–14)

10/10 issues closed. Floor inversion complete for all read-only categories. Canonical retest run 3 (Apr 13): 93.4% routing, 62.3% quality.

M2b Test infrastructure ✅ COMPLETE (Apr 14–15)

5/5 issues closed. E2E + AAXT + CI integrated. Canonical retest automated.

M2c Conversational depth ✅ COMPLETE (Apr 16)

#950 (Five Pillars + grammar + anti-flattening + Identity anchoring), #951 (calendar + deadline context), #964 (ethics verification — surfaced #1002+#1003+#1004). Canonical retest run 5 post-#950 iter 2: 72.1% quality PASS, 95.1% routing.

M2d MUX Lifecycle MVP ✅ CLOSED (end-of-day May 3)

8 implementation issues shipped (#704/#714/#1030/#1031/#1032/#1033 + pre-work #1034/#1035; #1036 closed premise-invalid). May 2 audit-cascade restructure surfaced flattening risks; conceptual-integrity gate added to m2-structure.md; UI Lifecycle Verification Rubric v0.1 branched per Methodology-24 (May 10) for per-issue gate-close protocol.

M2e Integrations 🎯 GAMEPLANS WALKED (May 3)

PM dispositions captured: #790 already shipped; #869 relocated from M2d (audit-cascade ✅); #900 (LLM-gated completion in MVP per Q2, ~14h est); #1039/#1040 (split from #864, #1039 blocked by #1042 PRE-1039 cleanup); #1041 M2-WIRE-TRIAGE (resolves WIRE-* ambiguity that was previously a footnote).

Followup issues filed during M2e walkthrough: #1043 POST-MVP en-masse copy review; #1044 FOLLOWUP local-git query handler; #1045 POST-MVP Project Detail Activity tab.

M2f Security + Infrastructure (gated)

Per CEO directive: M2f doesn’t open until canonical retest meets/exceeds prior benchmarks. Run 5 (May 9) baseline at 68.3% with v2.4-pending recalibration; v2.4 (C=0 disambiguation per CXO May 10 memo) is the durable fix; (b) interim 2-dim auto-fail threshold is in force as the recalibration Lead Dev shipped May 9. M2f opens once Run 5+ meets the M2c 72.1% benchmark.

Items per v15.0 stay; gate methodology is canonical-retest-baseline.

M3 — Artifact Persistence (post-MVP-related candidates surfacing)

Theme + scope unchanged from v15.0. Post-MVP-related issues filed during M2 walkthrough surface as M3 candidates — #1037, #1043, #1045, #1046 (most filed during M2d/M2e walkthroughs) are M3-shaped artifact-persistence-adjacent work to be triaged at M2 closure.

Composting data model designed in M3; composting engine deferred to M4.

M4 — Trust + Learning

Theme + scope unchanged from v15.0.

Trust graduation MVP must be credible first step toward full model, not throwaway hack (PM emphasis from v15.0 — still load-bearing). #1032 Push-insight trust-stage gating (Stage 3+) is the first concrete trust-gated UX surface; the trust thresholds it depends on are M4 territory.

M5 — Distribution + Polish

Theme + scope unchanged from v15.0.

BYOC discovery thread now in flight (PDR-005 scoping outline distributed May 4 to PA + Architect + CXO):

#992 ETHICS-ACTIVATE Arc — CLOSED (Apr 30)

Multi-step arc Phase A → B → C → D → E → #1002/#1003 → #1004 → Phase F. Six calendar days from initial S1 floor-bypass finding (Apr 25) to Phase F flag-flip merged (Apr 30 deecc816). #992 closed properly via skill (8/8 ACs marked + closing comment with full Phase F evidence).

Key landings:

Methodology Corpus (compounding since Apr 26)

Methodology landings since v15.0 — all responsive to operational findings within hours-to-days:

Patterns

Methodologies

Rubrics

Discipline Norms (CLAUDE.md-codified)

Distribution Strategy: Bring Your Own Chat (Discovery in Flight)

Build sequence (Gall’s Law) — unchanged from v15.0:

  1. MCP server — standalone, runs via stdio transport
  2. MCPB packaging — bundle for Claude Desktop one-click install
  3. Claude Project template — persona/instructions for the hybrid approach
  4. MCP Apps — interactive HTML for artifact canvas

PDR-005 discovery thread opened May 4 (post-Ship-#040 publication inflection point). Six decision-rule questions in scoping outline; tier-placement question (PDR-005 Foundational vs PDR-201 Integration Patterns — PPM lean is foundational, CEO call); paired-document approach with eventual BYOC ADR (slot TBD when Architect files).

HOST 360 cohort surfacing: Architect §8.3 + PPM §8.3 independent convergence — “the strongest decision-debt signal in the cohort” (Apr 27).

Build the server cleanly enough that the packaging layer is swappable — anchor on the thin-wrapper-to-API model, not on MCP specifically.

Methodology Maintenance (Per CIO + HOST + PPM observation)

Methodology-as-product is an asset only while it’s a living practice. The mechanisms that keep it alive (v15.0 list + new since Apr 26):

Per v15.0 (still operating):

New since v15.0:

These don’t appear in the feature roadmap but are required for the differentiator stack to remain credible.

Sprint Summary

Sprint Theme Status
M0 Conversational Glue ✅ COMPLETE (v0.8.6, Mar 4)
M1 MVP Foundation ✅ COMPLETE (gate closed Apr 11)
M2 Conscious Floor + Action Handlers 🎯 MID-SPRINT (M2a/b/c/d-MVP closed; M2e gameplan; M2f gated)
M3 Artifact Persistence — (post-MVP-related candidates surfacing)
M4 Trust + Learning — (#1032 Push-insight trust-gating is first concrete touch surface)
M5 Distribution + Polish — (BYOC PDR-005 discovery thread in flight)

Timeline (Inchworm, Not Calendar — Sequence Statements, Not Deadlines)

These are sequence statements. We are time lords. Each phase complete before the next begins.

Recent (April–May 2026)

Estimated forward sequence

Closures and Revisions Since v15.0

Issues filed since v15.0 (selected — full list in git): #1002, #1003, #1004 (#992 arc); #1005-1008 (audit_transparency cluster); #1011-1019 (Architect Phase 1 lens); #1027-1029 (dead-code follow-ups); #1030-1036 (M2d MUX restructure); #1037-1046 (M2d/M2e walkthrough followups); #1049 (weekly docs audit); #1064 (rubric recalibration trigger).

Roadmap Refresh Cadence (New, per Docs May 4 proposal + chesterton’s-fence preservation)

Hybrid mechanism (per PPM+CEO May 10):

  1. Trigger-based — refresh within 2 weeks of any sub-epic closure or major artifact landing (PDR ratification, ADR v1.0 ratification, multi-step arc closure). The natural cadence for product-roadmap deltas.
  2. Workstream-review-line-item — PPM compiles M2/M3 deltas as part of the workstream memo to Exec; the writeup itself becomes the source for the roadmap edit.
  3. Weekly docs audit retained per chesterton’s fence — Docs’s standing weekly audit continues as backstop; surfaces staleness if hybrid mechanism above misses something.
  4. Session-start hook — future enhancement (lower priority); PPM session-start could show roadmap last-updated days alongside briefing staleness counter. Not gating the hybrid above.

Change Log