What problem does it solve? Designing caching in GO CMS is error-prone: modules can couple to physical stores, keys and tags get mixed up, and cross-store invalidation breaks cache coherence. This Skill defines the architectural model and implementation rules so cache changes stay correct and consistent. ## Core Features & Use Cases - Layered cache model: Separates physical stores (Redis, filesystem, memory) from module-local aliases, deterministic keys, and semantic dependency tags. - Cross-store invalidation: Defines a runtime-owned coherence mechanism so one mutation invalidates dependent entries across aliases and stores without exposing the global cache manager to modules. - Read-through caching and testing: Prescribes a typed Remember helper, TTL-plus-tag policy, fail-open observability, and behavior-driven tests for hits, misses, expiry, and cross-store invalidation. - Use Case: When adding caching to a resource repository, use this Skill to choose the alias, key format, dependency tags, TTL, and the mutation paths that must trigger invalidation, then write the required behavioral tests. ## Quick Start Ask the AI to design or review a cache implementation for a GO CMS module following the go-cms-cache rules, specifying the store, alias, key, tags, TTL, and invalidation paths.