pom-core

Run one end-to-end intake funnel pass for one or more stakeholder source documents — research → ingest (score) → disposition → route (if routed) → optional Discovery gate → optional promote. Stops at human gates.

Stand up a new product end-to-end: spin up the product directory, form the first pod, draft an initial runway plan. Stops if any prerequisite is missing.

Run a full portfolio review — structural validation, portfolio audit (stuck UCs, stale runways, orphan ADRs), and ethics review of recent Discovery items. Read-only. Produces a stakeholder-ready summary.

Top-level router for the Product Operating Model toolkit — shows what's available, recommends the next step based on the current POM repo's state, and routes to the appropriate skill or command.

Independently grades a Discovery item against the four-question gate, without seeing the original grading. Provides a second-opinion verdict (✅/🟡/❌ per question) with rationale. Read-only. Use when a Discovery decision is high-stakes, controversial, or near the promote threshold.

Independently reviews a Discovery item's Q4 (ethical / responsible) grading, pushing back on weakly evidenced ✅ marks. Industry-aware — uses the generic Q4 framing plus any installed industry overlay framing. Read-only. Use after /pom-discovery-gate for high-stakes items, or in /pom-review for batch re-audits.

Scans the entire POM portfolio and surfaces structural and flow-level issues: stuck UCs, stale runways, orphan ADRs, products without pods, broken disposition→Discovery→Backlog chains. Returns findings grouped by category. Read-only. Used by /pom-review to produce stakeholder-ready summaries.

Provides a single role's framing on a Discovery item or product decision — Product Manager (viability), Product Designer (desirability), Technical Leader (feasibility), Scrum Master (process), or Product Owner (acceptance signals). Invoke with role=pm|pd|tl|sm|po as the first prompt argument. Reads the target DISC/PB/product files, never writes. Use when a Discovery item needs a Trio-style review without convening four humans.

Researches a raw stakeholder source (PDF, deck, markdown, transcript) and produces a structured pre-UC dossier — problem statement, stakeholders, outcomes, evidence, what's already true, open questions — ready for WSJF scoring by pom-wsjf-scorer or pom-ingest-use-case. Read-only on the POM repo. Use proactively when ingesting external materials before scoring.

Use when initializing a new Product Operating Model (POM) repository from scratch — scaffolds the intake funnel, product/platform/enabling layers, and a WSJF rubric (interactive or industry preset: insurance, fintech, healthcare, retail, generic).

Use when a precedent-setting decision needs to be appended to an enabling concern's decision log (AI ethics, security, etc.) — exceptions, threshold-setting, standards-level deviations, or conflicts with authoritative policy. Append-only. Not for UC dispositions or runway ADRs.

Use when a Discovery item needs to be walked through the four-question gate (desirable / viable / feasible / ethical) — produces a structured assessment with evidence, blockers, promotion-readiness verdict, and critical-path unblock sequence. Append-only on shaping log.

Use when a scored UC needs a formal kill/park/route/split decision. Writes a DISP file with required fields, paste-ready stakeholder summary, updates UC status. Append-only — reversals create new DISP that supersedes the old.

Use when a product is ready to stand up a new pod (typically its first, when a PB item is ready to pull). Copies pods/_template/, fills roster (SM, PO, 2-5 members), enforces 2-7 composition rule, documents surface area and cadence.