What problem does it solve? Manually auditing a codebase for SEO issues requires choosing among dozens of disciplines (technical SEO, schema markup, sitemaps, hreflang, Core Web Vitals) and then editing source files one finding at a time. Valhalla removes that decision and execution burden by perceiving the project, deciding which SEO disciplines apply, and applying every relevant fix directly to the source in a single run. ## Core Features & Use Cases - Autonomous full-project SEO pass: Runs a six-phase pipeline (perceive, decide, analyze, plan, apply, report) that ends with one reviewable diff on a feat/seo-pass branch, never the default branch. - Deterministic stack detection: Uses a stdlib-only scanner (seo_inventory.py) to identify the framework, routes, existing SEO surface, and business type before any edits. - 26 bundled discipline skills: Drives internal knowledge for technical SEO, content quality, schema.org, sitemaps, images, GEO/AI search, local, e-commerce, hreflang, SXO, clustering, and drift monitoring, loading each only when relevant. - Safety guardrails: Never fabricates facts (addresses, reviews, authorship), never touches secrets or build output, avoids deprecated schema types like HowTo, and uses INP rather than the retired FID metric. - Use Case: Point it at a Next.js marketing site and it detects missing metadata, generates appropriate JSON-LD, fixes sitemap and heading issues, and hands back a single branch diff to merge or revert. ## Quick Start Ask the assistant to run /valhalla on the current project to audit the codebase and fix all applicable SEO issues on a new branch.