h1-mobile-checklisted

# H1 Mobile Break Checker Production-ready checker for Japanese H1/H2 headings that detects mid-particle splits, dangling tail characters, and over-long headings at 375px mobile width. Combines static analysis (character count + grammar) with dynamic Playwright rendering (actual line-box measurement) and proposes `<br />` insertions at natural bunsetsu boundaries. ## What it does 1. **Static H1/H2 extraction** from `.tsx`, `.jsx`, `.html`, `.md`, `.mdx` files or live URLs 2. **Length check**: warns when H1 exceeds 15 characters (will likely wrap on 375px) 3. **Dynamic measurement**: renders the page at 375x667 with Playwright, reads `getBoundingClientRect()` per line, flags tail lines under 3 characters 4. **Particle-split detection**: flags Critical when a line starts with a Japanese particle (`を`, `が`, `に`, `で`, `と`, `は`, `から`, `まで`, `より`, `へ`, `の`) 5. **Verb-stem-split detection**: flags Warning when a verb is broken between stem and ending (e.g. `もたら` + `す`) 6. **`<br />` suggestion**: analyzes bunsetsu boundaries and proposes the optimal insertion point so both halves are visually balanced and grammatically intact ## When to use - Before every deploy that touches H1/H2 copy - After translating English headings into Japanese - When auditing existing blog/marketing pages for mobile UX regressions - In CI as a gate against H1 length regressions ## Critical rules (enforced) Per the netsujo monorepo `CLAUDE.md` headline rules: 1. **No 1-2 character tail lines** — the