Mk Preview

mk:preview

Generates visual artifacts — markdown and self-contained HTML — for explaining code, drawing diagrams, building slide decks, and visualizing diffs. Rendering a plan as HTML belongs to mk:visual-plan.

No live server. No Python. Pure markdown + HTML + bash file operations. Output writes to tasks/plans/{active-plan}/visuals/ when an active plan exists; otherwise falls back to tasks/visuals/.

When to Use

• A reader needs to understand an unfamiliar code path, protocol, or architecture
• A diagram (flowchart, sequence, ER) would communicate the topology faster than prose
• Step-by-step content benefits from slides over a single long page
• A pre-PR diff review needs visual KPIs, file map, and change cards
• Stakeholders prefer a self-contained HTML page they can open in a browser without a server

For rendering a plan directory as a shareable HTML page, use mk:visual-plan.

Default (No Arguments)

If invoked with no arguments, present available operations via stop and ask the user in chat. Header: "Preview Operation". Question: "What would you like to do?".

Recommended default: --explain or --html --explain.

Usage

Markdown Generation

• the preview skill --explain <topic> — visual explanation (ASCII + Mermaid + concepts)
• the preview skill --diagram <topic> — focused diagram (ASCII + Mermaid)
• the preview skill --slides <topic> — presentation slides (one concept per slide)
• the preview skill --ascii <topic> — ASCII-only diagram (terminal-friendly)

HTML Generation

• the preview skill --html --explain <topic> — self-contained HTML explanation
• the preview skill --html --diagram <topic> — HTML diagram with zoom/pan
• the preview skill --html --slides <topic> — magazine-quality slide deck

Analytical Modes

• the preview skill --html --diff [ref] — visualize a git diff. Default ref = main. Accepts branch, commit, range, PR number.

--ascii does not combine with --html (terminal-only by design). To render a plan as HTML, use mk:visual-plan.

Argument Resolution

Priority order:

1. --html flag detected → set HTML output mode
2. Generation flag detected (--explain, --diagram, --slides, --ascii) → load references/generation-modes.md
3. HTML-only flag (--diff) → implies --html; load references/analytical-modes.md
4. Topic missing → ask user via stop and ask the user in chat
5. Topic present → continue

Topic-to-slug:

• Lowercase the topic
• Replace spaces and special chars with hyphens
• Remove non-alphanumeric except hyphens
• Collapse multiple hyphens to single
• Trim leading/trailing hyphens
• Truncate at 80 chars on a word boundary

Title placeholder {topic} uses the original input in title case, not the slug.

Multiple flags: if more than one generation flag is supplied, use the first; the rest become part of the topic string.

Output Path Lifecycle

session-state/active-plan exists?
  yes → value is absolute path?
          yes → {value}/visuals/
          no  → tasks/plans/{value}/visuals/   (treated as slug)
  no  → tasks/visuals/

Detail and the bash detection snippet live in references/generation-modes.md → "Step 1 — Resolve output path". The fallback path is logged on stderr (warn:) so silent path mismatches surface immediately.

Error Handling

Reference Loading

Every mode reads its references BEFORE writing the output file.

Templates in assets/ are out-of-band — they are not auto-Read; the agent reads the template only when generating the matching mode.

Composes With

• mk:ui-design-system — palette and typography selection (assets/colors.csv 160 rows, assets/typography.csv 73 rows). HTML modes vary palette per run.
• mk:frontend-design — anti-slop forbidden patterns. Cited via ../frontend-design/references/anti-slop-directives.md, not duplicated.
• mk:web-to-markdown — for users who want to view a generated markdown file in a browser, no server bundled here.
• mk:visual-plan — the owner of plan-as-HTML rendering; route plan-render requests there.

Gotchas

• Mermaid .node class collision — Mermaid.js uses .node internally. Page-level .node CSS leaks into diagrams. Use .ve-card or any non-.node class on cards.
• Mandatory theme toggle — every HTML artifact MUST include the light/dark toggle button as the first child of <body> per references/html-design-rules.md. Missing toggle = incomplete output.
• Mermaid theme is static at load — switching the page theme does not re-skin Mermaid SVG internals; the diagram color palette is read once at init. Document this; do not pretend otherwise.
• Output path falls back silently — when session-state/active-plan is missing or unreadable, output writes to tasks/visuals/. Log the fallback on stderr so users notice; do not raise.
• Slug truncation at 80 chars — topics get lowercased, hyphenated, stripped of non-alphanumerics, and truncated at word boundary. Title placeholder uses the original input in title case (not slug). Mismatch silently produces wrong filenames.
• Topic strings are HTML-untrusted — interpolating --explain "</title><script>alert(1)</script>" MUST render as literal text. HTML-entity-encode < > & " ' in element and attribute contexts. See references/html-design-rules.md for context-by-context encoding rules.
• Browser auto-open is conditional — headless / SSH / WSL environments cannot open a browser. The shell snippet detects these cases and prints the path instead of invoking open/xdg-open.

Workflow Position

• Phase: on-demand
• Follows: nothing required
• Precedes: nothing required
• Common pairings: invoked after a researcher report or planner output to communicate findings; invoked before a review meeting to explain a code path or diagram an architecture.

Composes Into

• mk:cook may invoke --explain to communicate a complex implementation phase
• mk:visual-plan owns plan-as-HTML rendering; mk:plan-ceo-review CRITIQUES plans — neither overlaps this skill's generic explain/diagram/slides/diff output