Test Case Review Checklist

QA leads and senior engineers run this before a new or modified test case is merged into the regression suite. Catches unclear assertions, weak test data, missing traceability, and compliance gaps that turn into flaky CI or audit findings later.

6 sections 21 steps Collects data
1

Test Case Identity & Linkage

  1. Verify the title names the scenario, not the screen
    • Good: "Checkout rejects expired credit card with field-level error." Weak: "Checkout test 4." Title should let a reviewer skim the suite report and know what failed without opening the case.

  2. Confirm the test ID follows the suite naming convention
    • Match the existing pattern in the repo (e.g., TC-CHECKOUT-042, billing.subscription.upgrade.spec.ts). Inconsistent IDs break dashboards that group by feature area and confuse Jira links.

  3. Link the test case to its Jira ticket or user story
    • Paste the Jira/Linear key into the test metadata so traceability reports populate automatically. Tests with no upstream link are the first thing a SOC 2 auditor flags during change-management evidence review.

    Collects text
2

Preconditions & Test Data

  1. Validate stated preconditions and environment requirements
    • Preconditions should name the environment (staging, ephemeral preview, local Docker), feature-flag state, seeded user role, and any required service stubs. Implicit "works on my machine" preconditions are the leading cause of flakes when the test runs in CI.

  2. Classify the test data source
    • Synthetic fixtures are the default. Anonymized snapshots are acceptable when realistic distributions matter. Production-clone data triggers a PII review — never let raw prod records sit in a test repo.

    Collects list
  3. Confirm PII has been scrubbed from production-clone fixtures
    • Run the data through the team's anonymization script and spot-check email, name, address, payment, and free-text fields. Document the script version used. GDPR Article 5 and HIPAA minimum-necessary both apply if the clone leaves the production VPC.

  4. Document upstream and downstream test dependencies
    • List any other test cases that must run first (login fixtures, seed jobs) and any external services the test calls (payment sandbox, email mock, third-party API). Hidden test ordering is the second-leading cause of flakes after timing.

3

Steps & Expected Results

  1. Walk through each step for clarity and ordering
    • Each step should be a single user-observable action. "Click Submit and verify the toast and check the database" is three steps, not one. Reviewers should be able to execute the case manually from the written steps without reading the code.

  2. Verify expected results are stated for every step
    • An expected result of "page loads" is too vague — name the element, status code, or DB row that proves success. Steps without explicit expectations rot into smoke tests that pass while the feature is broken.

  3. Check that assertions use specific matchers
    • Prefer toEqual on a known shape over toBeTruthy. Prefer toHaveText('$49.00') over toBeVisible. Loose assertions are how a regression that swaps the price field with the SKU field still passes the test.

  4. Cover at least one negative or edge-case path
    • Happy-path-only suites are why bugs reach production. Confirm the case (or its sibling cases) covers invalid input, expired sessions, network failure, or boundary values relevant to the feature.

4

Post-Conditions & Cleanup

  1. Review the post-execution system state description
    • The case should state what the database, queue, and external integrations look like after a successful run. Without this, the next test inherits unknown state and you get order-dependent failures in CI.

  2. Confirm cleanup reverts database and filesystem changes
    • Use transaction rollback, fixture teardown hooks (afterEach), or explicit DELETE — never rely on "the next test will overwrite it." Leftover rows are the reason Friday's test run passes and Monday's fails.

  3. Verify external resources are released after teardown
    • S3 uploads, Stripe test customers, SendGrid mock recipients, and ephemeral k8s namespaces all leak quota if not deleted. The test should clean these up even when an assertion fails — wrap teardown in finally.

5

Traceability & Compliance

  1. Map the test case back to its acceptance criteria
    • Open the linked Jira ticket and confirm each AC bullet is covered by at least one assertion in this case (or another linked case). Unmapped ACs are how features ship half-tested.

  2. Determine whether the test exercises a regulated data path
    • If the code under test handles PHI (HIPAA), cardholder data (PCI DSS), EU resident data (GDPR), or controls referenced in the SOC 2 audit (access, change management, encryption), it falls in scope and needs an explicit control mapping.

    Collects list Collects text
  3. Cross-reference the regulatory control evidenced by the test
    • Tag the case with the specific control ID (e.g., SOC 2 CC7.2, HIPAA §164.312(b), PCI DSS 10.2). Vanta/Drata pull these tags directly into the audit evidence package — saves hours during the SOC 2 fieldwork window.

6

Peer Review & Sign-Off

  1. Run a spelling and grammar pass on the case description
    • Test descriptions land in customer-facing audit reports and onboarding docs. Typos in test names also break grep — "Checkuot" doesn't match anyone's search.

  2. Commit the test case to the version-controlled suite
    • Open the PR with the linked ticket key in the title, request review from CODEOWNERS for the affected directory, and confirm CI is green before requesting sign-off. Don't merge with a rerun-the-flake exception.

  3. Record the peer reviewer's decision
    • Reviewer must be someone other than the author — a self-approved test case is a SOC 2 segregation-of-duties finding. Capture both the decision and any rework notes so they survive after the PR is merged or closed.

    Collects list Collects paragraph
  4. File a rework ticket for the rejected test case
    • Open a Jira ticket linked to the original story, paste the reviewer notes verbatim, and assign back to the test author. Don't let rejected cases live in a draft branch — they get rebased into oblivion and the gap reappears in the next release.

Use this template

Copy it to your account, customize the steps, and run it with your team in minutes.


Sections 6
Steps 21
Category Software Development
Price Free to start
Need a different process

Browse hundreds of free templates across every team and industry.

Back to template library

Run Test Case Review Checklist with your team

Customize the steps, assign roles, set a schedule, and keep a complete record for every run.