skills.rest
haryelramalho
haryelramalhoCOMMUNITY
Explore all skills by this author →
Website𝕏@haryel_devGitHub

testing-boss

Review and repair tests for brittleness, flakiness, and mock-driven confidence.

Updated May 28, 2026
Search Tags:#testing#llm evaluation#flaky tests#mocking#test review#agentic workflows#ci
One-click install
npx skills add https://github.com/haryelramalho/skills --skill testing-boss-haryelramalho
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: testing-boss
Source: https://github.com/haryelramalho/skills/tree/main/skills/curated/testing-boss
Command: npx skills add https://github.com/haryelramalho/skills --skill testing-boss-haryelramalho

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you design, review, and repair tests that actually protect behavior instead of merely boosting coverage or mirroring implementation details. It gives you a clear doctrine for choosing the right test layer, avoiding brittle selectors, reducing flakiness, and preventing mock-driven confidence in both human-written and AI-generated tests.

Core Features & Use Cases

  • Test placement and boundary decisions: Decide whether a case belongs in unit, integration, route, or end-to-end testing based on the invariant being protected.
  • Test quality review: Identify brittle selectors, hidden shared state, static sleeps, weak assertions, over-mocking, and other failure modes.
  • AI test generation guardrails: Apply mandatory gates for agent-written tests, including invariant-first setup, real execution, and negative companions.
  • CI and flakiness discipline: Handle flaky tests with quarantine, ownership, deterministic design, and quality probes like mutation testing.
  • LLM and agent evaluation: Build stronger evals for RAG, tool use, and agent workflows using oracle ladders and outcome-based scoring.
  • Use case: When a new feature, failing test, or flaky suite needs diagnosis, this Skill helps you choose the right fix and the right test shape before touching code.

Quick Start

Ask for a test review or new test plan by naming the invariant, the owning layer, and the canonical suite, then let the Skill guide the rest.

Frequently Asked Questions about testing-boss

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

FAQPage Schema
How do I fix flaky tests in my CI pipeline?▼

Improve test quality by identifying brittle, flaky, and mock-driven tests through an invariant-first review. This process applies observable-behavior assertions and negative companions to ensure your tests catch real bugs without mirroring implementation details.

How do I review AI-generated tests for quality?▼

Review AI-generated tests by applying mandatory guardrails like invariant-first setup, real-system validation, and negative companions. This ensures agent-written tests protect actual behavior rather than merely boosting coverage with brittle, mock-driven assertions.

What is the best way to choose between unit and integration test layers?▼

Choose the right test layer by identifying the invariant being protected first. This test placement doctrine decides whether a case belongs in unit, integration, route, or end-to-end testing based on the specific behavior boundary.

How do I build stronger LLM evaluation workflows for RAG and agents?▼

Build stronger LLM evaluation workflows by using oracle ladders and outcome-based scoring for RAG, tool use, and agentic workflows. This approach evaluates actual observable behavior instead of relying on brittle implementation details.

When should I quarantine a failing test instead of fixing it immediately?▼

Quarantine a failing test when diagnosing flaky suites that disrupt CI pipelines. This test triage process enforces ownership and deterministic design rules, allowing you to isolate unstable tests while applying proper fixes.

Related Skills

Explore More Software Engineering Skills→
mepukamepuka

effect-layers-services

Define service interfaces and bind live/test implementations through layers.

Community
Intermediate
ArdurAIArdurAI
1

god-testing-mastery

Covers unit, integration, E2E, performance, chaos, fuzz, and mutation testing methodologies.

Official
Advanced
KrismithReddy12KrismithReddy12

health

Calculate a weighted 0-10 code quality score from linters and test runners.

Community
Advanced
rudironsonirudironsoni

dotnet-test-quality

Analyze .NET test quality using Coverlet, ReportGenerator, and Stryker.NET.

Community
Intermediate