Skip to main content

​
When to Use This

Use this guide when your app changes and existing tests need updates: new UI structure, altered API responses, or refined acceptance criteria.

​
Options at a Glance

  • Regenerate impacted tests via plans (recommended for broad changes)
  • Apply auto-healing suggestions (selectors, waits, fixtures)
  • Manually adjust specific tests and rerun quickly

​
Typical Workflows

​
1) Let the Agent Regenerate What Changed

Regenerates plans and tests based on current code and PRD. 
Refresh tests for the updated checkout and auth flows.
Steps:
  1. Bootstrap in diff scope
  2. Recreate test plans (frontend/backend as needed)
  3. Generate and execute
testsprite_bootstrap_tests({
  localPort: 5173,
  type: "frontend",
  projectPath: "/absolute/path/to/project",
  testScope: "diff"
})

testsprite_generate_frontend_test_plan({ projectPath: "/absolute/path/to/project", needLogin: true })
// or
testsprite_generate_backend_test_plan({ projectPath: "/absolute/path/to/project" })

testsprite_generate_code_and_execute({
  projectName: "your-project",
  projectPath: "/absolute/path/to/project",
  testIds: [],
  additionalInstruction: "Focus on updated modules"
})

​
2) Use Auto-Healing Suggestions

For selector drift, timing, or test-data issues, apply safe healing.
  • Review recommendations in the report under testsprite_tests/
  • Accept proposed edits in your IDE when prompted
  • Rerun to validate

​
3) Manually Edit and Rerun a Subset

When only a few tests need tweaks. 
Run TC005 and TC006 only after I fix the locators.

testsprite_generate_code_and_execute({
  projectName: "your-project",
  projectPath: "/absolute/path/to/project",
  testIds: ["TC005", "TC006"],
  additionalInstruction: "Use stable role-based selectors"
})

​
Rerun Previously Generated Tests

Use rerun to validate fixes without re-planning. 
testsprite_rerun_tests({
  projectPath: "/absolute/path/to/project"
})

​
Best Practices

  • Keep PRD and acceptance criteria in sync with product decisions
  • Prefer semantic selectors and explicit waits
  • Commit changes so diff-based regeneration stays accurate
  • Use small targeted reruns for fast iteration