Generate tests for untested functions and edge cases, then verify they pass.

Process

Identify context:

• If code changes are involved: run jj diff -s first to see changed files; then use jj diff -- path to restrict to specific files/directories

• If the user specified specific files or paths, focus on those

• Understand which functions/modules need test coverage

Detect test framework:

• Invoke code-analyze-project skill to detect test framework and testing patterns

• Check for existing test files and naming conventions

• Identify test helper functions and fixtures in use

Identify untested code paths:

• Use Grep to find functions without corresponding tests

• Check for edge cases, error paths, boundary conditions

• Identify critical paths that lack coverage

Research and analyze:

• Research testing best practices for the language/framework (use WebSearch/WebFetch)

• Identify critical paths, edge cases, and error handling scenarios

Generate test code:

• Match existing test conventions (naming, structure, fixtures)

• Cover happy path, edge cases, and error scenarios

• Include proper assertions and test organization

Run the generated tests:

• Invoke test skill to run the new tests

• Analyze any failures

Fix test failures:

• Identify root cause of failures

• Fix test code or generated code as appropriate

• Re-run tests to verify

Stop Condition

• If test generation or fixing fails twice, stop and ask for guidance.

Output

• Test Framework Detected

• Untested Functions/Paths Identified

• Generated Tests (with file locations and coverage summary)

• Test Results (from running tests)

• Failures Fixed (if any, with explanations)

• Verification - All new tests pass