How to create custom test rules
Extend Aucert's AI-generated tests with custom rules tailored to your application's specific requirements.
warning
Custom test rules are an advanced feature planned for Phase 2. This page describes the intended design. Core test generation and execution work today without custom rules.
Planned capabilities
Custom assertions
Define app-specific pass/fail criteria that the AI uses during analysis:
aucert.config.yaml (planned)
rules:
assertions:
- name: "Brand colors only"
description: "Primary buttons must use #1E88E5 background"
scope: [user-flows]
- name: "No placeholder text"
description: "No 'Lorem ipsum' or 'TODO' text in any screen"
scope: [navigation, user-flows]
Skip rules
Exclude specific screens or flows from testing:
aucert.config.yaml (planned)
rules:
skip:
- screen: "DebugSettingsActivity"
reason: "Internal debug screen, not user-facing"
- flow: "AdminPanel*"
reason: "Admin flows tested separately"
Platform-specific overrides
Different behavior for different form factors:
aucert.config.yaml (planned)
rules:
overrides:
- when: { screen_width_dp: ">= 600" }
then: { confidence_threshold: 0.90 }
reason: "Tablet layouts need higher scrutiny"
- when: { os_version: ">= 14" }
then: { skip: ["LegacyPermissionFlow"] }
reason: "Android 14+ uses new permission model"
Current alternatives
While custom rules are in development, you can achieve similar results:
| Need | Current workaround |
|---|---|
| Skip screens | Remove the screen from your test APK's navigation graph |
| Adjust confidence | Set testing.confidence_threshold globally in config |
| Focus on specific flows | Use testing.scope to limit to navigation or user-flows |
| Platform-specific testing | Build separate APKs with different configs |
What's next
- Configuration — Available configuration options
- Interpret results — Understanding test output