Vision
Auto-detect severity keywords and render them as visually distinct cards with color and icons. Critical rules pop off the page — no more scanning walls of text.
Detection Patterns
- Critical: MUST, NEVER, CRITICAL, FORBIDDEN
- Warning: SHOULD, WARNING, IMPORTANT, AVOID
- Info: NOTE, TIP, HINT, REMEMBER
Rendering
- Distinct card style per severity: border color, background tint, icon
- Works on inline text (paragraph containing keyword) and blockquotes
- Should feel native to each theme (clean/editorial/terminal each style differently)
Open Questions
- Detection scope: whole paragraph? Just the sentence? Blockquotes only?
- Should users be able to define custom keywords/patterns?
- Toggle on/off per file or globally?
- How aggressive should detection be? (avoid false positives in code blocks)
- How does this interact with GitHub-style
> [!NOTE] / > [!WARNING] admonitions?
Why This Matters
Instant visual value with zero user effort. Unique for markdown previewers. Especially powerful for CLAUDE.md and rule-heavy docs.
Vision
Auto-detect severity keywords and render them as visually distinct cards with color and icons. Critical rules pop off the page — no more scanning walls of text.
Detection Patterns
Rendering
Open Questions
> [!NOTE]/> [!WARNING]admonitions?Why This Matters
Instant visual value with zero user effort. Unique for markdown previewers. Especially powerful for CLAUDE.md and rule-heavy docs.