Release v5.0.0 · joenandez/spectre

Breaking

/spectre:ux_spec renamed to /spectre:ux — update aliases, scripts, and muscle memory

/spectre:prototype command — generates self-contained, clickable HTML prototypes. Modes: post-ux (renders complete spec, high-fi), flows-only, scope-only, --explore (pre-scope), and standalone
/spectre:ux ↔ /spectre:prototype integration — /ux inline-offers /prototype at the spec-review prompt; /prototype post-ux mode surfaces filled assumptions in chat with one-reply offer to write them back into ux.md; iteration loop handles spec↔prototype contradictions explicitly
/spectre:scope adds anti-scope as a first-class boundary (the problem we're intentionally not solving — distinct from OUT)
/spectre:scope adds Load-Bearing Assumptions section to scope.md so future-you knows when to come back
/spectre:ux adds User Segments section and segment-aware flow branching
/spectre:ux adds expanded State Vocabulary: visual / data / form / selection / sync / network states

/spectre:scope now reads existing scope.md / task_context.md / project knowledge before re-asking
/spectre:scope grounds its hypothesis with a single fast call (@finder or grep) before the question dump
/spectre:scope detects AskUserQuestion and routes Path A (inline) vs Path B (file) — matches /create_plan pattern
/spectre:scope marks questions as (blocking) vs (optional) so users can skim
/spectre:ux dispatches @spectre:patterns to find existing screens/components before proposing flows
/spectre:ux now leads with a proposed take ("Flows approved?") rather than open-ended approval ask

/spectre:ux (was /spectre:ux_spec) was reading task_summary.md / prd.md instead of the canonical concepts/scope.md

/spectre:scope Step 5 (Light Technical Context) — technical research is /spectre:plan's responsibility