What problem does it solve? Shipped copy drifts from a project's writing standards over time, accumulating AI-writing tells, banned words, and tone mismatches that manual review misses. This Skill systematically scans prose against a single content guide and reports every violation with a suggested rewrite. ## Core Features & Use Cases - Scoped auditing: Accepts a file path or one of many named scopes (readme, registry, case-studies, chat, nav, footer, palette, mcp, cli, and more) that map to specific data files, page modules, and string literals across the repo. - Three severity levels: Classifies findings as Banned (hard-ban words, em dashes, Title Case, first person where prohibited), Rationed (density-checked words, American spellings), or Judgment (rhythm uniformity, rule-of-three stacks, register mismatches). - Actionable reports: Each finding includes the file path and line, the offending text, the rule it breaks, and a suggested rewrite, ending with per-severity counts and the strongest finding. - Use Case: Ask to audit the chat scope before a release to catch promotional tone or banned words in the site chat's greeting, starter chips, and guardrail notices that no build validator covers. ## Quick Start Ask the assistant to run a content audit on the readme scope and report any violations of content-design.md with suggested rewrites.
