What problem does it solve? Hardcoded strings, magic numbers, and duplicated literals scattered across a TypeScript codebase create maintenance risk, stringly-typed control flow, and inconsistent naming. This Skill audits files, folders, or entire repos to find every loose literal and produces a scored report with a phased remediation plan. ## Core Features & Use Cases - Literal Detection and Classification: Finds magic numbers, statuses, events, routes, storage keys, headers, error codes, regexes, and repeated UI copy, clustering duplicates and near-duplicates. - Scored Ratings and Findings Table: Produces 0-100 scores across Literal Hygiene, Type Safety of Literals, DRY Centralization, and Enforcement, plus a severity-ranked findings table with proposed fixes. - Phased Remediation Plan: Generates MAGIC_LITERALS_REPORT.md and MAGIC_LITERALS_TASKS.md with lint guardrails, domain-based centralization, codemod-friendly replacement sequences, and CI lock-in steps. - Use Case: Point it at a frontend module full of raw status strings in reducers and hardcoded API routes, and receive a typed const-map refactor plan with union types derived via as const. ## Quick Start Ask the assistant to run the literal-hunter audit on a specific file, folder, or the entire repo and produce the findings report with centralization recommendations.
