Status: accepted · ADR-10 · Filed 2026-04-17
Decision
Expand ptype vocabulary beyond ‘application’ to include: research (academic papers, literature reviews), travel (trip planning, itineraries, bookings), home-improvement (product research, contractor decisions, long-running action plans), and leave the list open-ended for future additions. Each ptype gets its own prism-<ptype>.md overlay template that flavors the universal methodology for the domain. The ontology (Plan, Spec, Decision, TODO, Journal, Retro, Note, Memory) is confirmed domain-neutral; only texture differs. The persona system (Donna, Lola) becomes the cultural overlay — same identity, ptype-informed domain fluency.
Rationale
The noun/verb audit confirmed the domain model is universal. If Prism only works for code, it is a dev tool. If it works across research, travel, and home improvement, it is a general-purpose methodology for knowledge work with AI agents — a meaningfully larger product. The travel ptype specifically is the stress test: if session lifecycle, ADRs, and semantic_recall feel natural for trip planning, the generalization holds. If they feel forced, the code-specific assumptions are exposed and can be refactored. Expanding ptype is a low-cost extension point (write a new template overlay) with high strategic value (product scope).
Alternatives Considered
Keep Prism as code-project-only — simpler, but loses the general-purpose methodology positioning. Build separate tools per domain — duplicates infrastructure, loses cross-domain memory. Add a single ‘research’ ptype — insufficient granularity for travel vs research vs home-improvement.
