How We Test Tools
Each tool release follows a baseline quality workflow. The goal is to keep behavior deterministic where possible, and clearly labeled where output is probabilistic or format-based.
1. Structure and Routing Checks
- Every tool page must map route slug, widget mode, and guide slug consistently.
- Each tool page must include Introduction, How It Works, and FAQ sections.
- Each tool page must include a Reference Sample.
2. Runtime Handler Coverage
- Mode-based tools are checked against config and runtime handler definitions.
- Missing handler or missing config for any mode fails QA.
- Build and Astro type checks run before release.
3. Reference Sample Policy
- Exact output: deterministic tools where expected output should match exactly.
- Format output: tools with runtime/time/random/model variance.
- Reference examples are reviewed when input UX changes (for example, moving from line-config to selects).
4. Regression Process
- Run tool QA audit for full page coverage.
- Run Astro checks.
- Run production build.
- Review high-impact tools manually when behavior changes.
Related: Known Limitations, Data Handling Summary, Editorial Policy.