What problem does it solve? After running a GEO audit, teams face a backlog of findings (missing schema markup, robots.txt AI bot rules, canonical tags, llms.txt, freshness signals) that require careful, framework-aware code changes. This Skill automates the safe application of those fixes while gating risky changes behind explicit confirmation. ## Core Features & Use Cases - Safety-tiered fix application: Classifies findings as SAFE, MODERATE, or DANGEROUS, auto-applying safe fixes and requiring confirmation for dangerous ones like robots.txt AI policy changes. - Framework-aware patching: Detects Astro, Next.js, and Hugo stacks and targets the correct layout, config, or public files for schema.org JSON-LD, canonical tags, sitemaps, and llms.txt. - Verification and rollback: Runs build verification with exit-code checks, per-finding rollback on failure, adversarial review, and deduplication against seo-fix output to prevent double-application. - Use Case: After running geo-audit on an Astro site and getting a score of 41/100, run geo-fix to automatically add Organization JSON-LD and AI bot allow rules, then receive a report showing the estimated score improvement. ## Quick Start Apply the safe fixes from my latest geo-audit report and show me which dangerous fixes need my confirmation.