Portfolio PDLC — Wire

Outcome

A folder that satisfies the portfolio workspace contract: charter, workflow definition, one card per significant investment with honest (not aspirational) frontmatter, and a first generated board that already tells leadership something true. Start with what exists — map, don't invent.

Inputs

Point me at the material: a folder tree, exports (CSV/Jira/Align), decks, meeting notes — or name the initiatives conversationally. Also useful: who sponsors the portfolio, and any existing workflow/stage language the org already uses.

Workflow

1. Inventory before structure. Sweep the provided material and list candidate initiatives with whatever stage/owner/size evidence exists. Present the inventory before creating anything — merging duplicates and spotting "one initiative wearing three names" happens here.
2. Instantiate the workspace from skills/portfolio-pdlc/templates/ (charter, pdlc, topology, AGENTS.md). Keep the default lifecycle stages unless the org has real stage language — then map theirs onto the default in the pdlc.md table (the narrative aliases column exists for exactly this).
3. Have the two wiring dialogues (with the human, or record explicit defaults if running headless and flag them in the charter):
   • Flow boundaries: where does tracking start (idea? business ask?) and end (outcome measured → operational roadmap)?
   • Altitude: scale-up scoring rule (size + strategic + collaboration, 1–3 each) or enterprise rule (Baseline Gate + 2-of-4 constraint tests)? Score every candidate; 3–4 → Tier-2 awareness lane, 5–6 → flag for conversation, 7–9 → Tier-1.
4. Create the cards. One initiatives/<slug>/initiative.md per investment from the template. Honesty rules:
   • stage = where the evidence says it is, not where anyone wishes; note doubts inline.
   • stage_entered = best evidence (doc dates, commit history, "roughly March" → first of month, noted as approximate).
   • orientation: mixed unless the source material genuinely steers on outcomes — the assess skill upgrades it, never the wire.
   • Leave gaps visible (no outcome_hypothesis yet is a finding, not an embarrassment).
5. Backfill history if it exists. Prior stage transitions worth having (from status reports or tracker history) may be seeded into flow-log.csv — this is the ONE sanctioned hand-touch of that file, done chronologically, before the first board run.
6. Generate the first board: python3 <this-repo>/skills/portfolio-pdlc/scripts/portfolio_board.py <ws>.
7. Write the wiring brief (reviews/YYYY-MM-DD-wiring-brief.md): what got mapped, what got merged, altitude calls made, the board's first-read headlines (busy board? risk balance? oldest items?), and the open questions for the sponsors.

Rules

• Map, don't invent: every card traces to source material or a named human statement. Fabricated detail poisons every later assess.
• Don't fix while wiring. Activity-framed titles, missing hypotheses, stale stages — record them as they are; improving cards is portfolio-pdlc-strengthen's job on later cycles.
• Tiering is a proposal until a sponsor confirms it; say so in the wiring brief.
• A busy first board is information, not failure — the minibook's three causes (too much in flight / over-centralized / factions) go in the brief as hypotheses, not accusations.

Quality Gates

• Workspace passes the board script with zero contract violations.
• Every card names its source material.
• Flow boundaries and the altitude rule are written into the charter (or flagged as defaulted-pending-sponsor).
• The wiring brief states the board's first three headlines in sponsor language.

References

• skills/portfolio-pdlc/references/workspace-contract.md, templates/
• Umbrella: portfolio-pdlc (operating loop starts after wiring)