qa-test-cases

---

name: qa-test-cases description: Generate BDD/Gherkin test cases from acceptance criteria with ISTQB techniques. Use when writing test specs. allowed-tools: Read, Glob, Grep, Write($JAAN_OUTPUTS_DIR/qa/**), Task, WebSearch, Edit(jaan-to/config/settings.yaml) argument-hint: [acceptance-criteria | prd-path | jira-id | (interactive)] [--contract backend-api-contract-path] license: PROPRIETARY

qa-test-cases

Generate production-ready BDD/Gherkin test cases from acceptance criteria.

Context Files

• $JAAN_LEARN_DIR/jaan-to-qa-test-cases.learn.md - Past lessons (loaded in Pre-Execution)
• $JAAN_TEMPLATES_DIR/jaan-to-qa-test-cases.template.md - BDD/Gherkin template
• $JAAN_TEMPLATES_DIR/jaan-to-qa-test-cases-quality-checklist.template.md - Quality checklist template
• Research: $JAAN_OUTPUTS_DIR/research/50-qa-test-cases.md - ISTQB standards, test design techniques
• ${CLAUDE_PLUGIN_ROOT}/docs/extending/language-protocol.md - Language resolution protocol

Input

Acceptance Criteria Source: $ARGUMENTS

Input modes:

1. Direct text: Paste acceptance criteria directly
2. PRD path: Path to PRD file (from /jaan-to:pm-prd-write output)
3. Jira ID: Story ID (if Jira MCP available)
4. Interactive: Empty arguments triggers wizard
5. --contract {path} — Optional OpenAPI contract path (from /jaan-to:backend-api-contract output)

Contract parsing: If --contract present, extract path first, validate file exists and contains openapi: or swagger: key, then parse remaining arguments by modes 1-4 above.

IMPORTANT: The input above is your starting point. Determine mode and proceed accordingly.

---

Pre-Execution Protocol

MANDATORY — Read and execute ALL steps in: ${CLAUDE_PLUGIN_ROOT}/docs/extending/pre-execution-protocol.md Skill name: qa-test-cases Execute: Step 0 (Init Guard) → A (Load Lessons) → B (Resolve Template) → C (Offer Template Seeding)

Also read the comprehensive research document: $JAAN_OUTPUTS_DIR/research/50-qa-test-cases.md

This provides:

• ISTQB test case specification standards (Section 2)
• BDD/Gherkin format guidance (Section 2)
• Test design techniques: Equivalence Partitioning, BVA, Decision Tables (Section 3)
• Eight-step transformation workflow (Section 4)
• Edge case taxonomy - 5 priority categories (Section 5)
• Quality validation checklist (Section 6)
• AI failure mode mitigation patterns (Section 8)

If files do not exist, continue without them.

Language Settings

Read and apply language protocol: ${CLAUDE_PLUGIN_ROOT}/docs/extending/language-protocol.md Override field for this skill: language_qa-test-cases

---

PHASE 1: Analysis (Read-Only)

Step 1: Determine Input Mode and Extract AC

Check $ARGUMENTS to identify input mode:

Mode A - Direct Text Input: If $ARGUMENTS contains acceptance criteria text:

1. Parse the AC directly
2. Extract testable conditions
3. Proceed to Step 2

Mode B - PRD File Path: If $ARGUMENTS contains file path pattern (e.g., ".md" or "$JAAN_OUTPUTS_DIR/"):

1. Use Read tool to open the file
2. Locate "Acceptance Criteria" or "User Stories" section
3. Extract all AC statements
4. Preview extracted AC:

   EXTRACTED ACCEPTANCE CRITERIA
   ──────────────────────────────
   1. {ac_1}
   2. {ac_2}
   3. {ac_3}
   

5. Ask: "Use these acceptance criteria? [y/edit]"

Mode C - Jira MCP Integration: If $ARGUMENTS matches pattern "PROJ-123" or "JIRA-123":

1. Check if Jira MCP is available
2. If available: Fetch story → extract AC → preview
3. If unavailable: "Jira MCP not detected. Please paste AC or provide file path."

Mode D - Interactive Wizard: If $ARGUMENTS is empty:

1. Ask: "How would you like to provide acceptance criteria?"

   [1] Paste AC text directly [2] Provide PRD file path [3] Enter Jira story ID

2. Based on selection, route to appropriate mode

Ambiguity Resolution (from research Section 4):

If AC is vague, ask these questions before proceeding:

1. "What happens when required fields are empty?"
2. "What are the min/max boundaries for each input field?"
3. "What specific error messages should display for each failure?"
4. "What happens on system/network failure mid-operation?"
5. "Are there concurrent user scenarios to consider?"
6. "What permissions/roles are required for this action?"

Step 1.5: API Contract Extraction (if --contract provided)

When an API contract is provided:

1. Read the OpenAPI spec file
2. Extract: endpoints (paths + methods), request/response schemas, error codes (4xx/5xx), auth schemes
3. For each endpoint, auto-generate test scenario seeds:
   • Request body validation: minLength, maxLength, enum, required constraints from schema
   • Error response scenarios: one per documented 4xx/5xx status code
   • Auth boundary: unauthenticated access, expired token, insufficient scope
   • Pagination edge cases: empty page, last page, invalid cursor (if pagination parameters exist)
4. These seeds feed into Step 2 as additional acceptance criteria alongside the PRD/user stories

Reference: See ${CLAUDE_PLUGIN_ROOT}/docs/extending/openapi-integration-reference.md for contract discovery patterns.

Step 2: Apply Test Design Techniques

For each acceptance criterion, analyze using research Section 3 methodologies:

Reference: See ${CLAUDE_PLUGIN_ROOT}/docs/extending/qa-test-cases-reference.md section "Test Design Techniques" for Equivalence Partitioning, Boundary Value Analysis, Edge Case Categorization (5 priority categories), and Test Generation Ratio (minimum 10 tests per AC).

Step 3: Generate Test Inventory Summary

Calculate totals and show plan:

TEST GENERATION PLAN
────────────────────────────────────────
Acceptance Criteria: {n} criteria

Test Breakdown (per AC):
  - Positive tests: {n} × 3 = {total}
  - Negative tests: {n} × 3 = {total}
  - Boundary tests: {n} × 2 = {total}
  - Edge case tests: {n} × 2 = {total}

Total Test Cases: {grand_total}

Edge Case Distribution:
  - Empty/Null States: {count}
  - Boundary Values: {count}
  - Error Conditions: {count}
  - Concurrent Operations: {count}
  - State Transitions: {count}

Coverage Targets (Research Section 5):
  - Positive: 30%
  - Negative: 40%
  - Edge: 30%
  - Industry standard: 70-80% coverage

Ask: "Proceed with test case generation? [y/edit]"

---

HARD STOP - Human Review Check

Show complete plan before generating:

FINAL CONFIRMATION
──────────────────────────────────────
Source: {input_mode}
Acceptance Criteria: {n} criteria
Total Test Cases: {count}

Output Format: BDD/Gherkin (with ISTQB conversion notes)
Output Folder: $JAAN_OUTPUTS_DIR/qa/cases/{id}-{slug}/
Main File: {id}-test-cases-{slug}.md
Quality Checklist: {id}-test-cases-quality-checklist-{slug}.md

Process:
1. Generate BDD/Gherkin scenarios using research patterns
2. Apply concrete test data values (no placeholders)
3. Add @tags for traceability and filtering
4. Include ISTQB conversion notes (research Section 2)
5. Generate quality checklist (research Section 6)
6. Preview before writing

"Proceed with test case generation? [y/n]"

Do NOT proceed to Phase 2 without explicit approval.

---

PHASE 2: Generation (Write Phase)

Step 3.5: Generation Performance Strategy

Optimize test case generation throughput:

Batch BDD Processing: Group related ACs (2-3) into a single reasoning pass. Generate all scenarios for the batch before moving to the next group. This reduces context-switching overhead by batching related work.

Few-Shot Seeding: Pre-seed generation with 2-3 successful examples from $JAAN_LEARN_DIR/ if available. Past high-quality scenarios reduce hallucination and improve consistency.

Coverage-Driven Priority: Adjust generation depth based on existing test coverage:

• >90% coverage area: Reduce edge case depth (1 edge case per AC instead of 2)
• Untested area (<30%): Double boundary tests (4 per AC instead of 2)
• Critical path (auth, payments): Always generate full depth regardless of coverage

Incremental Generation: When re-running for updated ACs, preserve passing test cases and only regenerate for changed or new ACs. Compare AC text hashes to detect changes.

Step 4: Generate BDD/Gherkin Test Cases

Following research Section 2 patterns and Section 4 worked examples.

4.0 Declarative Gherkin Enforcement Rule

CRITICAL: All scenarios MUST use declarative style (describe BEHAVIOR, not UI actions).

BAD (Imperative -- NEVER use):

When I click the "Login" button
When I enter "[e-mail]" in the email field
When I scroll down to the footer
When I select "Option A" from the dropdown

GOOD (Declarative -- ALWAYS use):

When the user submits valid credentials
When the user requests password reset
When the user selects their preferred plan
When the user completes the checkout process

Rule: NEVER use imperative UI interaction steps. Describe BEHAVIOR, not UI actions. Steps should be understandable by non-technical stakeholders.

4.0.1 Scenario Structure Limits

• Steps per scenario: 3-5 (Given/When/Then combined). If exceeding 5, split into multiple scenarios.
• Scenarios per feature: 5-10. If exceeding 10, split into sub-features.
• Scenario Outline promotion: Use Scenario Outline with Examples tables when 3+ input combinations exist for the same behavior pattern.

Example Scenario Outline:

Scenario Outline: User login with various credential types
  Given a registered user with <credential_type>
  When the user submits valid credentials
  Then the user should be authenticated within 3 seconds

  Examples:
    | credential_type     |
    | email and password  |
    | social login        |
    | SSO token           |

4.0.2 Standardized Step Templates

Use these declarative patterns as starting points:

• Given a {entity} exists with {attribute} "{value}"
• Given the user is authenticated as {role}
• When the user {action} the {entity}
• When the user submits {valid/invalid} {data_type}
• Then the {entity} should have {attribute} "{value}"
• Then the user should see {feedback_type}
• Then the system should {expected_behavior}

4.1 Feature Header

@{test-type} @priority-{level} @REQ-{id}
Feature: {Feature Name}
  As a {role}
  I want {goal}
  So that {benefit}

4.2 Background Section

Common preconditions shared across scenarios:

  Background:
    Given {common_precondition_1 with concrete data}
    And {common_precondition_2}
    And {system_state}

4.3 Generate Scenarios (Research Section 4 pattern)

Positive Tests (3 per AC):

  @smoke @positive @priority-critical @REQ-{id}
  Scenario: {Main happy path}
    Given I am on the {page_name} page
    When I enter "{concrete_value}" in the {field_name} field
    And I enter "{concrete_value}" in the {field_name} field
    And I click the "{button_label}" button
    Then I should be redirected to {path} within {n} seconds
    And I should see "{exact_message}" in the {element}

  @regression @positive @priority-high @REQ-{id}
  Scenario: {Alternative happy path}
    Given {different_valid_state}
    When {valid_action_variation}
    Then {expected_outcome_with_threshold}

Negative Tests (3 per AC):

  @regression @negative @priority-high @REQ-{id}
  Scenario: {Invalid input scenario}
    Given {precondition}
    When I enter "{invalid_concrete_value}" in the {field} field
    And I click the "{button}" button
    Then I should see error "{exact_error_text}"
    And I should remain on the {page} page
    And {system_state_unchanged}

Boundary Tests (2 per AC):

  @boundary @edge-case @priority-medium @REQ-{id}
  Scenario: {Minimum boundary}
    Given {setup}
    When I enter "{min_value}" in the {field} field
    Then {expected_result_at_boundary}

  @boundary @edge-case @priority-medium @REQ-{id}
  Scenario: {Maximum boundary}
    Given {setup}
    When I enter "{max_value}" in the {field} field
    Then {expected_result_at_boundary}

Edge Case Tests (2 per AC from priority categories):

  @edge-case @empty-state @priority-low @REQ-{id}
  Scenario: {Empty/null handling}
    Given {setup}
    When I leave {field} field empty
    Then I should see validation error "{exact_text}"

  @edge-case @concurrent @priority-high @REQ-{id}
  Scenario: {Concurrent operation}
    Given {multi_user_state}
    When {user_1_action} and {user_2_action} occur simultaneously
    Then {system_prevents_race_condition}

4.4 Test Data Standards

Reference: See ${CLAUDE_PLUGIN_ROOT}/docs/extending/qa-test-cases-reference.md section "Test Data Standards" for concrete value examples, BAD/GOOD patterns, and standard test data values.

4.5 Tagging Strategy

Reference: See ${CLAUDE_PLUGIN_ROOT}/docs/extending/qa-test-cases-reference.md section "Tagging Strategy" for systematic tag application guidelines (@smoke, @regression, @priority, @REQ, etc.).

Step 5: Generate ISTQB Conversion Notes

Reference: See ${CLAUDE_PLUGIN_ROOT}/docs/extending/qa-test-cases-reference.md section "ISTQB Field Mapping" for the BDD-to-ISTQB field mapping table and example conversion from Gherkin to ISTQB format.

Step 6: Quality Validation (Research Section 6)

Before preview, validate against 10-point checklist:

Universal Checks:

• All tests map to an AC (Alignment)
• All steps unambiguous with specific elements (Clarity)
• All preconditions/data/results documented (Completeness)
• Expected results measurable with thresholds (Measurable)
• All test data explicit values, no placeholders (Test Data)
• Traceability tags present (@REQ-{id})
• Tests independent, no hidden dependencies (Independence)
• Tests reproducible by any tester (Reproducibility)
• Negative coverage included (30%+) (Negative Coverage)
• Edge coverage addressed (5 categories) (Edge Coverage)

AI Failure Mode Checks (Research Section 8):

• No vague steps ("properly", "correctly", "works")
• No missing preconditions
• No non-reproducible scenarios (relative dates, undefined entities)
• No over-specification (database/internal references)
• 30/40/30 distribution (positive/negative/edge)

If any check fails, fix before proceeding.

Step 7: Preview & Approval

7.1 Generate Output Metadata

source "${CLAUDE_PLUGIN_ROOT}/scripts/lib/id-generator.sh"
SUBDOMAIN_DIR="$JAAN_OUTPUTS_DIR/qa/cases"
mkdir -p "$SUBDOMAIN_DIR"
NEXT_ID=$(generate_next_id "$SUBDOMAIN_DIR")

Generate slug:

• Extract feature name from AC or user title
• Convert to lowercase-kebab-case
• Max 50 characters
• Example: "User Login Authentication" → "user-login-authentication"

7.2 Generate Executive Summary

Template:

Production-ready BDD/Gherkin test cases for {feature_name} covering {n} acceptance criteria with {total_tests} test scenarios. Includes {positive_count} positive tests, {negative_count} negative tests, {boundary_count} boundary tests, and {edge_count} edge case tests across 5 priority categories (empty/null, boundary, error, concurrent, state transition). All tests use concrete test data values and include traceability tags.

7.3 Show Preview

OUTPUT PREVIEW
──────────────────────────────────────
ID: {NEXT_ID}
Folder: $JAAN_OUTPUTS_DIR/qa/cases/{NEXT_ID}-{slug}/
Main: {NEXT_ID}-test-cases-{slug}.md
Checklist: {NEXT_ID}-test-cases-quality-checklist-{slug}.md

# Test Cases: {Feature Name}

## Executive Summary
{executive_summary}

## Metadata
- Acceptance Criteria: {n}
- Total Test Cases: {count}
- Coverage: Positive {pct}%, Negative {pct}%, Edge {pct}%

[Show first 3 scenarios as preview]

@smoke @positive @priority-critical @REQ-001
Scenario: Successful login with valid credentials
  Given I am on the login page
  When I enter "[e-mail]" in the email field
  And I enter "ValidP@ss123!" in the password field
  And I click the "Login" button
  Then I should be redirected to the dashboard within 3 seconds
  And I should see "Welcome, Test User" in the header

[Full output contains {total_tests} scenarios]

Ask: "Write these output files? [y/n]"

Step 8: Write Output Files

If approved:

8.1 Create Folder

OUTPUT_FOLDER="$JAAN_OUTPUTS_DIR/qa/cases/${NEXT_ID}-${slug}"
mkdir -p "$OUTPUT_FOLDER"

8.2 Write Main File

Path: $OUTPUT_FOLDER/${NEXT_ID}-test-cases-${slug}.md

Use template from: $JAAN_TEMPLATES_DIR/jaan-to-qa-test-cases.template.md

Fill sections:

• Title, Executive Summary
• Metadata table
• Acceptance Criteria Coverage table
• BDD/Gherkin Test Scenarios (all scenarios)
• ISTQB Conversion Notes
• Traceability Matrix
• Test Execution Guidelines
• Quality Checklist Reference
• Appendix

8.3 Write Quality Checklist File

Path: $OUTPUT_FOLDER/${NEXT_ID}-test-cases-quality-checklist-${slug}.md

Use template from: $JAAN_TEMPLATES_DIR/jaan-to-qa-test-cases-quality-checklist.template.md

Fill sections:

• 10-Point Peer Review Checklist
• Anti-Patterns to Reject
• Quality Scoring Rubric (100-point scale with 6 dimensions)
• Coverage Sufficiency Analysis

8.4 Update Index

source "${CLAUDE_PLUGIN_ROOT}/scripts/lib/index-updater.sh"
add_to_index \
  "$SUBDOMAIN_DIR/README.md" \
  "$NEXT_ID" \
  "${NEXT_ID}-${slug}" \
  "{Feature Name} Test Cases" \
  "{Executive Summary}"

8.5 Confirm Completion

✅ TEST CASES GENERATED
───────────────────────────────────
ID: {NEXT_ID}
Folder: $JAAN_OUTPUTS_DIR/qa/cases/{NEXT_ID}-{slug}/
Main: {NEXT_ID}-test-cases-{slug}.md
Checklist: {NEXT_ID}-test-cases-quality-checklist-{slug}.md
Index: Updated $JAAN_OUTPUTS_DIR/qa/cases/README.md

Total: {count} test cases
Coverage: {positive_pct}% positive, {negative_pct}% negative, {edge_pct}% edge

Step 9: Capture Feedback

Ask: "Any feedback on the test cases? [y/n]"

If yes:

"[1] Fix now [2] Learn [3] Both"

Option 1 - Fix now:

• Ask what to improve
• Apply feedback
• Re-validate (Step 6)
• Re-preview (Step 7)
• Re-write

Option 2 - Learn:

• Run: /jaan-to:learn-add qa-test-cases "{feedback}"

Option 3 - Both:

• Fix current output (Option 1)
• Save lesson (Option 2)

If no: Complete

---

Skill Alignment

• Two-phase workflow with HARD STOP for human approval
• Multi-stack support via tech.md detection
• Template-driven output structure
• Output to standardized $JAAN_OUTPUTS_DIR path

Definition of Done

• AC extracted and parsed
• Test design techniques applied (EP, BVA, edge cases)
• BDD/Gherkin scenarios generated
• Concrete test data values (no placeholders)
• Quality validation passed (10-point checklist)
• ISTQB conversion notes included
• Executive Summary generated
• Sequential ID generated
• Folder created: {id}-{slug}/
• Main file written: {id}-test-cases-{slug}.md
• Quality checklist written: {id}-test-cases-quality-checklist-{slug}.md
• Index updated
• If API contract provided, every documented endpoint has at least 1 test scenario
• User approved