Use when
Testing impact is the main question
Use this when the task is to check whether behavior changes are covered by tests and whether regression-sensitive paths need new verification.
Test Coverage & Regression Review Prompt
Public prompt page for reviewing test coverage gaps, missing scenarios, and regression risk after code or configuration changes.
Public prompt
Review test coverage and regression risk
Use this prompt when a code or configuration change needs a coverage review, missing-scenario check, or regression-risk assessment. It is a code-review path focused on testing impact, not a general architecture review.
Do not use for
Architecture or standards review
Use the architecture or standards-backed implementation paths when the dominant problem is system structure or official-guidance compliance.
Access
Public
This prompt is public. It should not receive a Pro badge unless it is later moved behind a member route.
Prompt setup
Where to use this prompt
Use this as a runtime review prompt inside the code-review workflow. Keep stable team rules in the instruction layer and provide the current change, tests, config, logs, and coverage output as runtime material.
ChatGPT
Projects, GPTs, Skills, files/sources, Custom Instructions, Memory, and runtime prompts.
Open mapping →
Claude
Projects, Project Instructions, Project Knowledge, Skills, and runtime chat input.
Open mapping →
Gemini
Gems, Gem instructions, Knowledge files, Memory, Connected Apps, AI Studio, Gemini API, Files API, File Search, and Vertex AI.
Open mapping →
Inputs required
The prompt is designed to fail closed when evidence is missing. Provide the materials below before using it.
Change context
Goal, constraints, changed files, implementation excerpts, config, environment details, and affected behavior.
Test evidence
Relevant test files, fixtures, mocks, test data, logs, coverage output, and known regression-sensitive paths.
Copy-ready prompt
Paste the current materials after the final START line.
Run the TEST COVERAGE AND REGRESSION REVIEW on the materials below.
Use this review only for testing, coverage, regression risk, and verification impact.
Do not perform architecture review or standards-backed implementation review unless needed to explain test impact.
If the supplied materials are insufficient for a finding, write INSUFFICIENT_EVIDENCE.
OUTPUT EXACTLY in this order:
A) CONTEXT SNAPSHOT
B) TEST BASIS & SCOPE
C) COVERAGE FINDINGS
D) REGRESSION RISK FINDINGS
E) CHANGESET PLAN
F) CONFIDENCE
REVIEW RULES:
- Base findings only on the provided materials and on authoritative sources when supplied or when browsing is explicitly allowed.
- Identify the behavior under test, impacted components, interfaces, configuration, data/setup requirements, and test evidence available in the materials.
- Check whether the provided tests cover positive paths, negative paths, edge/boundary cases, error paths, authorization/authentication paths, integration points, and regression-sensitive behavior.
- Check whether assertions verify observable behavior and expected results, not only execution without failure.
- Check whether fixtures, mocks, test data, environment setup, and cleanup are consistent with the implementation changes.
- Check whether implementation changes require updates to unit, integration, API, end-to-end, security, or regression tests.
- Do not invent missing tests, coverage numbers, execution results, logs, defects, or pass/fail status.
- Mark every unsupported claim as NOT VERIFIED.
CHANGESET PLAN RULES:
- Propose only test/code/config/documentation changes that are directly supported by the provided materials.
- Prefer minimal, file-specific remediation steps.
- Separate required test additions from optional test improvements.
- Include exact missing test scenarios when they are supported by the provided materials.
- If test evidence is missing, state what material is required to complete the review.
MATERIALS (goal, constraints, implementation excerpts, test files, config, logs, coverage output, and optional authoritative sources or permission to browse) START: