---
argument-hint: "[testCommand]"
description: Fix the tests your task execution broke.
---

Fix all failing tests in this repository.

Use tools for debugging. You are encouraged to write your own specialized tools to research, analyze, and debug
failing tests. You are not restricted to the stack of the current project to create your own
specialized tools.

Execute this command to run the tests:

<test-command>
$1
</test-command>

If the test command is empty, read CLAUDE.md for test execution instructions.

⚠️ **CRITICAL INTEGRITY REQUIREMENT** ⚠️
You MUST fix the actual bugs in the source code. Green tests are worthless if achieved through cheating.

**This is CHEATING (absolutely forbidden):**
- Skipping tests with conditionals
- Modifying test assertions to pass
- Adding test-environment-specific code to source
- Disabling or commenting out tests
- ANY workaround that doesn't fix the real bug

**This is THE RIGHT WAY:**
- Find the root cause in the source code
- Fix the actual bug
- Ensure tests pass because the code truly works

**Process:**
1. Run all tests to identify failures
2. Fix EVERY failing test iteratively
3. Verify all tests pass legitimately

DO NOT STOP after fixing some tests - fix ALL of them.

Remember: The entire point of tests is to ensure code robustness. If you cheat in ANY way, the tests become meaningless and I cannot trust that the code actually works.

If you get stuck and cannot fix a test properly, ask for help rather than resorting to shortcuts.