AI Prompt for Test Coverage Gaps

Find risky untested code with this AI prompt: coverage gap analysis ranked by likelihood times impact, plus targeted tests for the riskiest gaps.

The prompt

Role: You are a test strategist who finds the untested code that is most likely to hide bugs.

Context: Language: [LANGUAGE], framework: [FRAMEWORK]. Module: [CODE]. Current coverage: [COVERAGE_INFO].

Task: 1) Analyze the code and list untested or under-tested paths: branches, error handlers, edge conditions, concurrency paths. 2) Rank them by RISK (likelihood × impact), not by line count — a tested happy path with untested error handling is riskier than untested trivial code. 3) Write tests for the top 5 riskiest gaps. 4) Identify code that is untestable as written and suggest the minimal refactor to make it testable.

Format: Gap list ranked by risk (gap | risk | why), then the tests, then the testability refactor suggestions.

Ask the AI to distinguish untestable-by-design code from merely-untested code, and to propose the minimal refactor for the former. Coverage you cannot achieve by writing more tests needs a design change, not more effort.

Constraints: Do not chase 100% coverage blindly — prioritize by risk. Flag tests that would be brittle or low-value so I can skip them deliberately.

Variables to fill in

Replace each [variable] in the prompt with your own details:

[LANGUAGE] [FRAMEWORK] [CODE] [COVERAGE_INFO]

Example input

LANGUAGE=Python, CODE=payment processing module, COVERAGE=72%, FRAMEWORK=pytest, FOCUS=untested branches in refund logic

How to use this prompt

Paste the coverage report summary if you have it — it focuses the analysis on real gaps. Point at the riskiest module; 100% coverage everywhere is not the goal.

Tips for the best results

Fix the top-3 risks first; diminishing returns kick in fast after that.
Ask which gaps are untestable by design — those need refactoring, not more test effort.
Re-run quarterly; coverage rots as features are added.