What problem does it solve? PPTX files generated from HTML slide decks via python-pptx silently drift from the source: content overflows into footers, italic emphasis disappears, hero slides lose vertical centering, and fonts fall back to system defaults. This Skill provides a repeatable five-step workflow to detect, diagnose, and fix these regressions. ## Core Features & Use Cases - Ground-truth extraction: Dump every shape's position, size, and per-run typography from a .pptx into JSON for comparison against the HTML source. - Severity-ranked audit table: Classify issues as critical (footer overlap, off-canvas), high (broken hierarchy), medium (italic/font loss), or low, with root-cause analysis. - Footer-rail + cursor-flow re-export: Rebuild the deck using a Cursor primitive that refuses to cross the content rail, budget-centered hero slides, and explicit italic propagation. - Geometry-based verification: Run a post-export checker that asserts rail and canvas invariants and exits nonzero on violations. - Use Case: You have deck.html and deck.pptx side by side and the PPTX shows text bleeding into the footer with missing italics. Run the audit to get a severity table, re-export with the layout discipline, and verify zero rail violations. ## Quick Start Ask the assistant to audit the exported deck.pptx against its source deck.html and re-export it with footer-rail layout discipline until verification passes.
