guardian

Write tests that validate production code without mutating it.

Updated Mar 26, 2026

Search Tags:#regression #guardian #production-code #tests #test-authoring #senior-sdet

Ready to run

npx skills add https://github.com/benediktms/overmind --skill guardian-benediktms

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: guardian
Source: https://github.com/benediktms/overmind/tree/main/cli/claudecode-plugin/skills/guardian
Command: npx skills add https://github.com/benediktms/overmind --skill guardian-benediktms

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The guardian persona enforces a strict testing discipline: you write tests that expose bugs without touching production code, ensuring regressions are caught before merge.

Core Features & Use Cases

Write tests for production behavior without modifying source.
Detect regressions by adding failing tests before bug fixes.
Follow project conventions and avoid mocking the unit under test.

Quick Start

Provide guardian-compliant tests that demonstrate the required behavior without editing any production files.

Frequently Asked Questions about guardian

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

FAQPage Schema

How do I write tests to catch regressions without modifying production code?▼

To write tests for regressions without modifying production code, author tests that validate existing behavior and expose bugs by following project conventions, covering happy paths and edge cases while avoiding mocking the unit under test.

What is the best way to add failing tests before fixing a bug?▼

The best way to add failing tests before fixing a bug is to author tests that demonstrate the required behavior, detecting regressions by validating production code strictly through test authoring without editing any production source files.

When do I need to avoid mocking the unit under test during test authoring?▼

You need to avoid mocking the unit under test during test authoring when applying strict testing disciplines to production modules, ensuring tests validate actual production behavior and catch regressions before merge without mutating source files.

Does this approach work for tasks labeled test-authoring and coverage across repository modules?▼

Yes, this test-authoring approach works for tasks labeled tests, coverage, or test-authoring across repository production modules by identifying and authoring tests that validate behavior without mocking the unit under test or touching production code.

What are the limitations of writing tests that never modify production source?▼

A limitation of writing tests that never modify production source is that you cannot fix the bug in the production code itself; you can only detect and expose it by authoring failing tests that demonstrate the required behavior and validate existing conventions.