What problem does it solve? cspell lint runs flag unknown words across a codebase, and fixing them ad hoc leads to bloated dictionaries, scattered suppressions, and inconsistent spelling. This Skill applies a strict decision tree that resolves each flagged word with the least-invasive correct fix, preferring rewording over config changes. ## Core Features & Use Cases - Reword-first resolution: Prioritizes rephrasing prose in American English (e.g., colour → color, datastore → data store) before touching any cspell configuration. - Scoped config management: Adds words, dictionaries, or regex patterns at the narrowest correct scope — inline comment, package-level cspell.json, or monorepo root. - Structured reporting: Outputs a per-word decision trace showing why rewording, dictionaries, or patterns were accepted or ruled out, then applies the fix. - Use Case: After running npx cspell on a monorepo and getting 30 unknown-word errors, use this Skill to triage each one — rewording prose, adding real technical terms to the right cspell.json, and suppressing one-off generated identifiers. ## Quick Start Run cspell on the repo and fix every unknown-word error it reports, preferring rewording over config changes.