What problem does it solve? Reading a raw technical SEO audit output (crawl/index issues, Core Web Vitals, rendering, schema, security findings) and deciding what to fix first is slow and error-prone. This Skill reads an existing audit's findings.json and turns the TECH-* findings into a prioritized, grounded fix list without inventing metrics. ## Core Features & Use Cases - Findings Interpretation: Pulls findings.json from a completed audit via the AIOS API and groups TECH-* findings into five areas: crawl/index, Core Web Vitals, rendering/JS, schema, and security/infra. - Prioritized Fix List: Ranks findings by severity then lowest per-check score, boosts by impact_usd when present, and flags low-confidence findings (< 0.6) as needing validation. - Grounding & Degradation Handling: Reports only measured values from the engine, labels deterministic-subset runs as degraded, and surfaces data gaps ([NEEDS: ...]) for human routing instead of fabricating data. - Use Case: An operator asks "read the technical part of audit abc-123" and receives a ranked fix list with quick wins, validation flags, and per-area counts, plus a next-step recommendation to re-audit after fixes. ## Quick Start Ask the assistant to interpret the technical findings of audit <audit-id> and produce a prioritized fix list.
