What are the three testing levels enforced by this skill?

Level 1 covers standard library and local tools; Level 2 involves project-specific binaries like Docker or Postgres; Level 3 includes real external network dependencies and SaaS APIs.

Why does this skill prohibit the use of mocking for external services?

It follows the principle that 'reality is the oracle,' preferring dependency injection with Protocols over mocks to ensure tests verify actual behavior against real systems.

Does this skill provide actionable feedback?

Yes, every architectural rejection includes the specific principle violated, the line numbers in question, and a code example of the correct architectural approach.

How does it handle SaaS APIs like Stripe or GitHub?

The skill enforces a strict jump from Level 1 to Level 3, as SaaS services cannot be reliably emulated locally (Level 2), requiring real-world workflow verification.

Python Architecture Reviewer

Name: Python Architecture Reviewer
Author: simonheimlicher

bysimonheimlicher

•

Security & Testing

Reviews Architecture Decision Records (ADRs) to ensure compliance with strict Python testing levels and modularity principles.

The Python Architecture Reviewer acts as a principal architect to audit ADRs and design proposals against a specialized testing-first methodology. It focuses on enforcing a rigid 'no-mocking' policy for external services, ensuring that Python projects utilize dependency injection and Protocol interfaces instead of fragile unit test mocks. By validating testing levels (1, 2, and 3) and verifying SaaS integration strategies, this skill helps developers build robust, reality-based architectures that are easier to test and maintain.

Key Features

01Enforcement of Protocol-based dependency injection patterns

021 GitHub stars

03Detection of prohibited mocking patterns in service boundaries

04Direct code-based architectural corrections for rejected designs

05Validation of SaaS integration jumps from Level 1 to Level 3

06Automated ADR auditing against standardized Python testing levels

Use Cases

01Reviewing new Architecture Decision Records before implementation

02Auditing project documentation for testing strategy alignment

03Refactoring legacy mock-heavy tests into Protocol-based injection

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add simonheimlicher/spx-claude reviewing-python-architecture

For use in Claude.ai and ChatGPT

Download Skill