test-fixing

Diagnose and fix failing tests by grouping errors and resolving root causes systematically.

1|Updated Aug 17, 2025
One-click install
npx skills add https://github.com/ratnesh-maurya/mdconverter --skill test-fixing-ratnesh-maurya
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: test-fixing
Source: https://github.com/ratnesh-maurya/mdconverter/tree/main/.claude/skills/test-fixing
Command: npx skills add https://github.com/ratnesh-maurya/mdconverter --skill test-fixing-ratnesh-maurya

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When a test suite breaks after a refactor or dependency change, developers face dozens of scattered failures with no clear starting point. This Skill turns that chaos into an ordered workflow by grouping failures by root cause and fixing them one group at a time. ## Core Features & Use Cases - Smart Error Grouping: Clusters failures by error type (ImportError, AttributeError, AssertionError), affected module, and root cause so related failures are fixed together. - Prioritized Fix Order: Resolves infrastructure issues (imports, dependencies, configuration) first, then API changes, then logic bugs, maximizing impact per fix. - Incremental Verification: Runs focused test subsets after each fix group before re-running the full suite to confirm no regressions. - Use Case: After renaming a module breaks 15 tests, the Skill groups the 8 ImportErrors, 5 AttributeErrors, and 2 AssertionErrors, fixes each group in order, and verifies the full suite passes. ## Quick Start Ask the assistant to run the test suite and systematically fix all failing tests by grouping similar errors together.

Frequently Asked Questions about test-fixing

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I fix multiple failing tests after a refactor?

Run the full test suite first, then group failures by error type and root cause such as renamed modules or changed function signatures. Fix the highest-impact group first, verify with a focused test run, and repeat until the full suite passes.

How to run only a subset of pytest tests?

Use pytest with a file path to run one test file, or the -k flag with a pattern to run matching tests, for example pytest -k "pattern" -v. This lets you verify one fix group without running the entire suite.

What order should I fix test failures in?

Fix infrastructure issues first, such as import errors, missing dependencies, and configuration problems. Then address API changes like renamed functions, and finally resolve logic issues like assertion failures and edge cases.

Why do tests fail after renaming a module or function?

Renames break every import and call site referencing the old name, producing ImportError or AttributeError failures across many tests. Use git diff to find the change, update all references, and re-run the affected test subset.

When should I not use grouped test fixing?

Avoid it when failures stem from a single isolated bug, since grouping adds overhead with no benefit. It is also less useful when tests fail due to external services or flaky environments rather than code changes.