What problem does it solve? Code reviews in the basic-memory and basic-memory-cloud repositories require consistent judgment about house style, architecture boundaries, and pre-merge risk, which generic review passes often miss. ## Core Features & Use Cases - Rubric-Based Review: Evaluates diffs for cognitive load, change propagation, knowledge duplication, accidental complexity, dependency direction, domain model distortion, and test oracle quality. - House Rule Enforcement: Explicitly checks for speculative getattr defaults, broad exception swallowing, hidden casts, dataclass/Pydantic misuse, and unclear async resource ownership. - Severity-Ranked Reporting: Outputs findings ordered by severity (high, medium, low) with file:line references, concrete evidence, and minimal fix suggestions. - Use Case: Before merging a pull request in basic-memory-cloud, run this skill to verify tenant/workspace isolation boundaries are respected and that risky changes include meaningful regression tests. ## Quick Start Ask the AI to review the current diff or named files using the code-review skill and report findings by severity.
