What problem does it solve? Drafting social and editorial content often leads to fabricated claims, unsourced statistics, and off-brand copy. This Skill produces unpublished, channel-specific drafts where every non-obvious claim is traceable to a named, linkable source, and publication decisions stay with the user. ## Core Features & Use Cases - Channel-specific drafting: Generates distinct drafts for X, LinkedIn, Reddit, Hacker News, Product Hunt, blogs, and newsletters with per-channel length and format targets. - Evidence ledger and fact-checking: Classifies every claim as verified fact, approved positioning, user opinion, inference, or unknown, with sources, caveats, and disclosure needs. - Approved file output and visuals: After explicit approval, saves drafts to sm_<slug>/posts.md with a Sources section, and can produce SVG/PNG hero and carousel visuals with a defined QA process. - Use Case: Provide a product launch summary and ask for LinkedIn and X drafts; receive unpublished drafts with an evidence ledger, disclosure notes, and claims flagged for approval before anything is saved. ## Quick Start Ask the assistant to draft a LinkedIn post and an X post from your supplied source material, keeping every statistic linked to a named source and leaving the drafts unpublished for your review.