architecture-review

Review proposed code changes against layered architecture rules and dependency direction constraints.

1|Updated Apr 8, 2026
One-click install
npx skills add https://github.com/voidful/Aixlarity --skill architecture-review-voidful
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: architecture-review
Source: https://github.com/voidful/Aixlarity/tree/main/.aixlarity/skills/architecture-review
Command: npx skills add https://github.com/voidful/Aixlarity --skill architecture-review-voidful

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Proposed changes can silently violate module boundaries, reverse dependency direction, or add unjustified dependencies to core crates. This Skill provides a structured checklist to catch layer violations, dependency cycles, and accidental coupling before they are merged. ## Core Features & Use Cases - Dependency Direction Enforcement: Verifies that aixlarity-cli depends on aixlarity-core one-way only, and that internal module arrows (tools → agent → prompt → app) are never reversed. - Layer Separation Checks: Confirms command loading, skill loading, and prompt assembly stay in separate layers without entanglement. - Trust and Testability Review: Ensures trust/sandbox decisions are explicit and core logic remains testable offline without API keys. - Use Case: Before merging a PR that adds a new tool requiring special permissions, run this review to confirm the change respects dependency direction, justifies any new aixlarity-core dependency, and keeps offline testability intact. ## Quick Start Review this proposed change against the architecture rules and flag any layer violations or unjustified new dependencies.

Frequently Asked Questions about architecture-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I review whether a code change violates layered architecture?

Check the change against five questions: does it respect dependency direction, does it add a justified dependency to the core crate, does it mix concerns across layers, can it be tested offline, and is the trust impact explicit. Any reversed dependency arrow is a violation.

How to prevent dependency cycles between crates in a Rust workspace?

Enforce a one-way dependency rule such as aixlarity-cli depending on aixlarity-core but never the reverse. During review, reject any change that makes the core crate import from the CLI crate or reverses internal module arrows.

When should I run an architecture review on a pull request?

Run it before any PR that adds a new module, significantly changes module boundaries, moves code between crates during refactoring, or adds a tool that might need special permissions. Routine small changes within one module usually do not need it.

What are the limitations of a checklist-based architecture review?

A checklist catches declared violations like reversed imports or mixed layers, but it cannot detect subtle coupling introduced through shared state or implicit contracts. It also depends on reviewers honestly answering whether new dependencies are justified.

Why must core logic be testable offline without API keys?

Offline testability keeps the core crate independent of external providers, so unit tests run deterministically in CI without secrets. A change that requires live API keys for core logic signals a layering problem where provider concerns leak into core.