What problem does it solve? Writing blog posts with AI usually produces generic, slop-filled text that does not sound like the author. This Skill produces long-form blog essays in Caio's personal voice, in both PT-BR and EN, with verified facts, controlled analogy and humor budgets, and a strict anti-slop pass for each language. ## Core Features & Use Cases - Structured writing pipeline: angle and thesis selection, mandatory fact verification against the repo and design notes, skeleton built from approved exemplars, PT draft, EN rewrite, and per-language anti-slop passes. - Anti-slop reference tables: PT-BR and EN tables of AI-writing patterns (inflated vocabulary, negative parallelism, rule of three, vague attribution) to detect and rewrite before publishing. - Exemplar-based structure: a few-shot bank of approved posts with named structural moves (M1-M9) plus rejected moves and the reasons they failed. - MDX and diagram contract: frontmatter format, slug pairing across locales, and rules for optional inline SVG diagram components registered in mdx-components.tsx. - Use Case: Ask for a post about a technical decision in your project; the Skill proposes 2-3 angles, verifies every fact against the repository, drafts the PT essay, rewrites it in EN, and runs the full checklist on both files. ## Quick Start Ask the assistant to write a blog post about a chosen topic using the write-article skill, pick one of the proposed angles, and let it produce the paired PT and EN MDX files.
