Show HN: SlideGauge – Static analyzer for Marp Markdown decks (github.com)

SlideGauge is a new zero-dependency Python static analyzer for Marp Markdown slide decks that produces structured, AI-agent-friendly feedback to improve slide quality. Built as a single ~1k-line Python file (stdlib only), it scores each slide (starting from 100) and emits JSON/SARIF/text reports, plus a stdio JSON protocol so coding agents can call it programmatically. It’s aimed at automating slide hygiene for agents and humans alike—giving clear, actionable fixes (e.g., “reduce by ~30 chars or split”) rather than vague warnings—and integrates easily into CI/CD and GitHub Code Scanning workflows. Technically, SlideGauge evaluates 11 rules across five buckets (content, layout, color, code, accessibility). Defaults include title ≤35 chars, content ≤350 chars (450 for exercises), ≤6 bullets, ≤15 lines, WCAG AA contrast ≥4.5:1, images must have alt text, and code length limits (≤10 simple / ≤5 complex lines). Rules deduct weighted points from slide scores, are configurable via JSON, can be disabled with in-slide comments, and the tool supports UUID-based caching for deterministic, fast re-analysis. It’s extensible through a simple rule registry decorator, outputs SARIF for scanners, and is distributed via pip/git with instant demo usage, making it practical for automated agents and developer pipelines.