Three additions to the title/outro colophon: 1. Optional avatar (`avatar: <url>` in frontmatter). On the outro, a large circular byline portrait sits at the top of the right-column stack. On the title, a smaller circular avatar pairs with the speaker name in a CSS grid: avatar in column 1, name in column 2, byline and context hang-indented under the name in column 2 so every text line shares the same left edge. Avatar center y-aligns with name center y (1px optical match). 2. Optional QR code on the outro (`qr: true` for current-URL, `qr: <url>` for a fixed canonical link). New <morkdeck-qr> Lit component renders client-side via @levischuck/tiny-qr + @levischuck/tiny-qr-svg — a dependency-free TypeScript encoder so the runtime bundle stays clean of Node-only modules. QR sits in the bottom-left of the outro grid below the "Thank you" wordmark with a "scan for these slides" label slotted in, left-aligned to match the rest of the slide's editorial flow. 3. Colophon markup unification. Dropped the .colophon-attribution wrapper. Avatar, speaker, byline, and context now live as siblings inside .colophon-author, with per-layout CSS arranging them: title uses a 2-column grid (avatar | text), outro uses a flex column (avatar above text). Same DOM, purposeful compositions. Density tier overrides for .colophon-speaker / .colophon-byline / .colophon-context margin: 0 with a 2-class selector (`.colophon .colophon-speaker`) — the slide-wide `[data-density="dense"] p` rule's margin-bottom was bleeding into the colophon text and shifting each line's optical position away from where align-self placed it. examples/showcase.md gains `avatar` + `qr: true` so the demo exercises both fields. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>