L-Ami-Fiduciaire/_bmad/tea/workflows/testarch/atdd/steps-c/step-02-generation-mode.md

name, description, outputFile, nextStepFile

name	description	outputFile	nextStepFile
step-02-generation-mode	Choose AI generation or recording mode	{test_artifacts}/atdd-checklist-{story_id}.md	./step-03-test-strategy.md

Step 2: Generation Mode Selection

STEP GOAL

Choose the appropriate generation mode for ATDD tests.

MANDATORY EXECUTION RULES

• 📖 Read the entire step file before acting
• ✅ Speak in {communication_language}

---

EXECUTION PROTOCOLS:

• 🎯 Follow the MANDATORY SEQUENCE exactly
• 💾 Record outputs before proceeding
• 📖 Load the next step only when instructed

CONTEXT BOUNDARIES:

• Available context: config, loaded artifacts, and knowledge fragments
• Focus: this step's goal only
• Limits: do not execute future steps
• Dependencies: prior steps' outputs (if any)

MANDATORY SEQUENCE

CRITICAL: Follow this sequence exactly. Do not skip, reorder, or improvise.

1. Default Mode: AI Generation

Use AI generation when:

• Acceptance criteria are clear
• Scenarios are standard (CRUD, auth, API, navigation)
• If {detected_stack} is backend: Always use AI generation (no browser recording needed)

Proceed directly to test strategy if this applies.

---

2. Optional Mode: Recording (Complex UI)

Skip this section entirely if {detected_stack} is backend. For backend projects, use AI generation from API documentation, OpenAPI/Swagger specs, or source code analysis instead.

If {detected_stack} is frontend or fullstack:

Use recording when UI interactions need live browser verification.

Tool selection based on config.tea_browser_automation:

If auto:

Note: ${timestamp} is a placeholder the agent should replace with a unique value (e.g., epoch seconds) for session isolation.

• Simple recording (snapshot selectors, capture structure): Use CLI
  • playwright-cli -s=tea-atdd-${timestamp} open <url> → playwright-cli -s=tea-atdd-${timestamp} snapshot → extract refs
• Complex recording (drag/drop, wizards, multi-step state): Use MCP
  • Full browser automation with rich tool semantics
• Fallback: If preferred tool unavailable, use the other; if neither, skip recording

If cli:

• Use Playwright CLI for all recording
• playwright-cli -s=tea-atdd-${timestamp} open <url>, snapshot, screenshot, click <ref>, etc.

If mcp:

• Use Playwright MCP tools for all recording (current behavior)
• Confirm MCP availability, record selectors and interactions

If none:

• Skip recording mode entirely, use AI generation from documentation

---

3. Confirm Mode

State the chosen mode and why. Then proceed.

---

4. Save Progress

Save this step's accumulated work to {outputFile}.

• If {outputFile} does not exist (first save), create it with YAML frontmatter:

  ---
  stepsCompleted: ['step-02-generation-mode']
  lastStep: 'step-02-generation-mode'
  lastSaved: '{date}'
  ---
  

  Then write this step's output below the frontmatter.

• If {outputFile} already exists, update:

  • Add 'step-02-generation-mode' to stepsCompleted array (only if not already present)
  • Set lastStep: 'step-02-generation-mode'
  • Set lastSaved: '{date}'
  • Append this step's output to the appropriate section.

Load next step: {nextStepFile}

🚨 SYSTEM SUCCESS/FAILURE METRICS:

✅ SUCCESS:

• Step completed in full with required outputs

❌ SYSTEM FAILURE:

• Skipped sequence steps or missing outputs Master Rule: Skipping steps is FORBIDDEN.