autorenew
PROMPT DIRECTORY 246 prompts · 11 categories

Software Testing Prompt Library

246 reusable prompts across 11 testing themes, with 3 guided workflows.

View GitHub
246 prompts 11 categories 3 workflows

246 matching prompts

BROWSE DIRECTORY

All Prompts

Browse by category and open a prompt by clicking its card.

20 prompts

28 prompts

36 prompts

16 prompts

35 prompts

11 prompts

18 prompts

25 prompts

16 prompts

23 prompts

18 prompts

Quick Start

  1. 1

    Choose the testing task

    Start from the testing problem you need to solve.

  2. 2

    Choose a Prompt version

    Use Standard by default and switch only for a specific structure.

  3. 3

    Prepare the input

    Provide requirements, APIs, user stories, logs, or defect details.

  4. 4

    Use an AI assistant

    Provide the Prompt, source material, and expected output format together.

  5. 5

    Review the output

    Check assumptions, omissions, risks, and unverifiable claims.

From input to testing artifact

These are structural examples; review real results against project context.

  • Requirements analysis before a feature release

    Standard

    InputLogin requirement text + product goals

    Why this versionComplete structure that yields a reviewable first draft

    Output excerpt

    • Scope summary: login, registration, and password recovery
    • Business rules: password policy, captcha expiry, device binding
    • Gaps: concurrent login and risk control are not specified
    • Risk priority: password recovery is high risk — verify first

    Human review points

    • Check rules against the latest requirements
    • Confirm no invented endpoints or fields
    View details →
  • API test design before the order service launch

    Standard

    InputOpenAPI document + auth method + environment notes

    Why this versionThe default structure covers the full API testing output

    Output excerpt

    • Contract coverage: create, query, and cancel order endpoints
    • Error paths: duplicate submit, timeout, invalid token
    • Assertions: status codes, field types, idempotency
    • Data strategy: placeholders and isolated environments

    Human review points

    • Smoke-test the assertions against a real environment
    • Confirm the auth method matches production
    View details →
  • UI automation approach for regression

    Standard

    InputCore regression paths + target browsers + CI setup

    Why this versionScope and strategy fit the default structure

    Output excerpt

    • Layering: smoke, core regression, and edge cases
    • Locators: prefer accessibility roles and stable identifiers
    • Run config: screenshots on failure, retry cap, timeout baseline

    Human review points

    • Verify locators stay stable on real pages
    • Check scripts do not depend on changing test data
    View details →
  • Standardized report for a production defect

    Standard

    InputSymptoms, log excerpt, draft reproduction steps

    Why this versionThe default structure matches the report fields

    Output excerpt

    • Title: payable order marked cancelled before the callback
    • Steps: five reproducible actions
    • Expected vs actual: status flow against the log timeline
    • Impact: volume of paid-but-cancelled orders

    Human review points

    • Reproduce once before submitting
    • Remove sensitive data from logs
    View details →
  • Test report assembly at iteration close

    Standard

    InputExecution results, defect list, release window

    Why this versionAggregates in the default structure for easy review

    Output excerpt

    • Execution summary: totals, pass rate, and blockers
    • Risk list: uncovered scope and known issues
    • Release advice: go / conditional go / hold with reasons

    Human review points

    • Data must come from real execution records
    • Advice must match actual defect severity
    View details →
  • Security verification points before launch

    Standard

    InputApp type, login system, data sensitivity

    Why this versionA checklist fits the plain default structure

    Output excerpt

    • Authentication: login, session, and permission boundaries
    • Inputs: injection, broken access, and file upload checks
    • Data: masking sensitive fields and log leakage checks

    Human review points

    • Use as a review reference; run security tools and manual checks for actual detection
    • Trim by risk level instead of applying everything
    View details →

AI-assisted Testing Flow Reference

Testing Workflows