review-clean-code

name	review-clean-code
description	Comprehensive Clean Code audit across eight rule families (comments C1-C5, environments E1-E2, functions F1-F4, naming N1-N7, architecture G1-G10, hygiene G11-G20, logic G21-G36, tests T1-T9). Partitioned — run all aspects or a chosen subset. Invoked by /ac:review.
metadata	{"origin":"ECC"}

Clean Code Review

A comprehensive Clean Code audit based on Robert C. Martin's heuristics. This skill is partitioned into eight aspects; each lives in its own part file and can be run independently or together.

Reference: rules/clean-code.md for the full catalog of rules.

Aspects

Aspect	Part file	Rules	Focus
comments	comments.md	C1-C5	Comment quality and accuracy
environments	environments.md	E1-E2	Build and test automation
functions	functions.md	F1-F4	Function design (args, output, flags)
naming	naming.md	N1-N7	Naming conventions
architecture	architecture.md	G1-G10	Code structure & macro architecture
hygiene	hygiene.md	G11-G20	Code hygiene & clarity
logic	logic.md	G21-G36	Logic & design decisions
tests	tests.md	T1-T9	Test code cleanliness & maintainability

Boundary: the tests aspect reviews existing test code hygiene only. Suite effectiveness, correctness, gaps, flakiness, and meaningful coverage belong to the top-level review-tests skill; select either or both independently as needed.

How to Run

1. Determine which aspects to run. When invoked by /ac:review, the requested aspect list is passed in (the hub asks the user "all or which parts?"). When invoked directly with no aspect specified, default to all eight.
2. For each selected aspect, load only that part file (skills/review-clean-code/<aspect>.md) and apply its rules to the target files. Loading parts on demand keeps context lean.
3. Each aspect cites specific Rule IDs and produces findings with minimal refactoring examples.
4. Aggregate all aspect outputs into a single Clean Code Audit Report, then return it (to /ac:review for consolidation, or directly to the user).

Output

One unified report grouped by aspect. For each finding: the violated Rule ID, the affected region, and a minimal refactoring example for the existing code under review, including existing tests when the tests aspect is selected.