What problem does it solve? Caching and compression misconfigurations fail silently: bandwidth or CDN costs rise with flat request volume, bundles get refetched on every navigation, and assets refuse to pick up new deploys, all without any error. Worse, the standard checks themselves can lie - a curl -I HEAD request can report a clean result that was never capable of detecting the problem. ## Core Features & Use Cases - False-clean detection: Explains why HEAD-based compression checks, no-cache without validators, and nginx add_header inheritance produce misleading results, and how to confirm with GET requests and wire-byte comparison. - Decision rules: Covers when immutable is safe (only with content-hashed URLs), how gzip_types literal matching against text/javascript breaks compression, and how ^~ prefix locations stop regex evaluation. - Use Case: Egress traffic jumps 3x after a frontend rewrite with flat request counts. Use this Skill to run GET-based header and byte-size checks at the CDN edge, identify a gzip_types mismatch and missing cache validators, and fix the nginx config. ## Quick Start Ask the AI to verify whether your site's static assets are actually being cached and compressed, using GET requests to compare response headers and wire bytes.
