code-hygiene

code-hygiene is a Codebase audits & code health skill for AI agents, published by cloudposse in the atmos repository on GitHub. Skillspace rates code-hygiene D (59/100) and it passed the safety scan; the rating is based on static analysis only.

Reviews code changes for architectural smells that mechanical checks (lint, tests, coverage) structurally can't catch: duplicated 'shared' abstractions, missing sentinel errors, generics that discard their own type info, self-aware nolint suppressions of mandated rules, business logic in the wrong layer, admitted-but-unshipped gaps, config/flags that validate but silently no-op or always error, features that are implemented but never wired up, and documentation/code mismatches in either direction (docs describing an unimplemented or since-changed feature, or a shipped feature with no docs at all). Distinct from the general code-review skill (bugs/security/perf) and from lint (mechanical/syntactic) — this catches CLAUDE.md architectural-mandate violations that need reading intent, not just syntax. Patch-scoped against origin/main by default; a full-repo sweep runs only on an explicit human request. Invoke on explicit requests like \"check for vibe coding\" / \"run code hygiene\" /…

Grade

D

Quality

59 / 100

Safety

Passed

GitHub stars

1,376

Unproven: rated by static analysis only, not yet compared head-to-head.

This rating describes the version of the repository we crawled; it may have changed since, so go by its current files.

Skillspace indexes and rates this skill; it does not host or endorse it. Run third-party skills at your own risk.