test-flakiness

Analyze CI logs to detect flaky tests and classify root causes.

Updated Mar 28, 2026
One-click install
npx skills add https://github.com/scomofo/dragon-forge --skill test-flakiness-scomofo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: test-flakiness
Source: https://github.com/scomofo/dragon-forge/tree/main/.agents/skills/test-flakiness
Command: npx skills add https://github.com/scomofo/dragon-forge --skill test-flakiness-scomofo

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Test flakiness causes unreliable CI results where the same test sometimes passes and sometimes fails without any code change, which wastes engineering time and makes real failures easier to overlook.

Core Features & Use Cases

  • Detect intermittent failures: Aggregates PASS/FAIL outcomes across multiple CI runs (or a specified CI log) to flag tests with mixed results.
  • Classify likely root causes: Infers common drivers such as timing/async, order dependency, random seeds, resource leaks, external state, floating point comparisons, and scene load races.
  • Recommend remediation: Suggests whether to quarantine immediately or to investigate and fix soon, and prepares updates to the regression-suite quarantine section and an optional flakiness report.
  • Use Case: When /regression-suite shows quarantined tests, run this skill to confirm which ones are genuinely flaky, explain why they fail intermittently, and propose concrete next actions.

Quick Start

Run the skill in registry mode to review already-flagged quarantined flaky tests and get recommended remediation steps for each one.

Frequently Asked Questions about test-flakiness

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

FAQPage Schema
How do I detect flaky tests from CI logs or JUnit XML artifacts?

To detect flaky tests from CI logs or JUnit XML artifacts, the skill aggregates pass/fail outcomes across multiple CI runs to flag tests with mixed results, identifying non-deterministic behavior. It analyzes plain-text logs or JUnit XML files to isolate intermittent failures.

What are common root causes of non-deterministic test failures in a regression suite?

Common root causes of non-deterministic test failures include timing or async issues, order dependency, random seeds, resource leaks, external state, floating point comparisons, and scene load races. The skill infers these drivers by inspecting static log and test patterns from your CI runs.

How do I manage quarantined tests and decide whether to fix or keep them isolated?

To manage quarantined tests, the skill reviews flagged tests in your regression suite and classifies flakiness severity to recommend whether to investigate and fix soon or quarantine immediately. It prepares updates to the quarantine section only on approval to maintain CI trust.

Does this test flakiness analysis work with Godot engine test result artifacts?

Yes, this test flakiness analysis works with Godot engine test result artifacts. It applies to workflows that scan Godot test result artifacts, including JUnit XML or plain-text logs, to aggregate pass/fail outcomes and classify non-deterministic test behavior.

Why does my test pass locally but fail intermittently in CI without any code change?

Intermittent CI failures without code changes happen because of non-deterministic test behavior like timing races, order dependency, or resource leaks. The skill aggregates multiple CI run outcomes to confirm which tests are genuinely flaky and explains why they fail.

When should I not use automated flakiness detection for my test suite?

You should not use automated flakiness detection when you lack multiple CI run logs or existing quarantine history, as the skill relies on aggregating pass/fail outcomes across runs to infer root causes and recommend remediation. Single run failures are insufficient for pattern detection.