What to build: Rewrite docs/adding-a-card.md for the policy reversal: full Startup coverage is now an explicit goal. Deck-driven imports remain the *prioritization* mechanism (a wanted deck's cards jump the queue) but are no longer the gate on what gets implemented.
Blocked by: None — can start immediately.
Status: fixed
2026-08-01: Rewrote the pipeline section of docs/adding-a-card.md — full Startup coverage is the explicit goal, pack manifests are the work queue (pack-major sweeps, backfill first), deck imports demoted to prioritization. Touched the two echoes of the old policy: CLAUDE.md's pipeline summary and CONTEXT.md's pack-manifest entry ("deck-driven pipeline" → "card pipeline").