Advanced Usage & Best Practices
This page explains how to get the most accurate, relevant, and useful test cases when using VibeTester.
1. Choosing the Right LLM
Section titled “1. Choosing the Right LLM”-
Not all models are optimised for speed. Some are slow and designed for deep analysis.
-
In most cases, a faster, cheaper model delivers results just as accurate and is better suited for test case generation.
2. Understanding the Three Levels of Context
Section titled “2. Understanding the Three Levels of Context”When generating test cases, the app passes information to the LLM in three layers:
-
System level – defined within the plugin and defines the required format (BDD, Manual, Generic).
-
Test case rules – ensures consistency across your team or organisation (naming, structure, terminology).
-
User story context – provides scenario-specific details such as acceptance criteria or unique conditions.
Use all three together for the most consistent results.
3. Writing Effective Jira Fields
Section titled “3. Writing Effective Jira Fields”The quality of generated cases depends heavily on the clarity of your Jira issues.
-
Write acceptance criteria in plain, specific language.
-
Include edge cases (invalid inputs, boundaries, error conditions).
-
Avoid vague terms like “System should be secure” instead use “System locks user account after 3 failed login attempts”
4. Reviewing AI-Generated Cases
Section titled “4. Reviewing AI-Generated Cases”AI accelerates test creation but human review remains essential.
Reviewer checklist:
-
Do steps align with acceptance criteria?
-
Are both positive and negative cases included?
-
Are expected results specific and measurable?
-
Does the case match the project’s test strategy?
5. VibeTester can be used to support Exploratory Testing
Section titled “5. VibeTester can be used to support Exploratory Testing”VibeTester can help teams generate exploratory test charters.
-
Provide a template in your test case rules to standardise test charter generation across your team.
-
Use in conjunction with Generic or Plain Text type test cases.
6. Limitations to Keep in Mind
Section titled “6. Limitations to Keep in Mind”-
AI may produce redundant or obvious cases — filter before saving.
-
Domain-specific details may still need manual editing.
-
The plugin does not replace risk-based or exploratory analysis.
9. Staying Connected
Section titled “9. Staying Connected”We are constantly improving VibeTester.
- If you’d like to suggest enhancements or request new integrations, please contact us at support@assertthat.com.