ia-writing-tests

What problem does it solve?

It prevents AI-generated or poorly written tests from being theater—tests that pass without proving real behavior, miss edge cases, or fail to expose silent failures and broken outcomes.

Core Features & Use Cases

• One behavior per test with clear, outcome-based naming to keep intent auditable.
• Coverage built from three sources (requirements, implemented behavior, and stated claims) to eliminate coverage gaps.
• Anti-pattern and failure-mode detection including mock-overuse, snapshot-only testing, circular assertions, and silent failure coverage.

Quick Start

Ask your agent to write or improve tests for the specific failing feature by deriving cases from the requirements, the diff, and your response claims, then ensuring each test asserts observable outcomes (including error and silent-failure paths).