Address three issues reported after overnight AutoForge runs: 1. ~193GB of .node files in %TEMP% from V8 compile caching 2. Stale npm artifact folders on drive root when %TEMP% fills up 3. PNG screenshot files left in project root by Playwright Changes: - Widen .node cleanup glob from ".78912*.node" to ".[0-9a-f]*.node" to match all V8 compile cache hex prefixes - Add "node-compile-cache" directory to temp cleanup patterns - Set NODE_COMPILE_CACHE="" in all subprocess environments (client.py, parallel_orchestrator.py, process_manager.py) to disable V8 compile caching at the source - Add cleanup_project_screenshots() to remove stale .png files from project directories (feature*-*.png, screenshot-*.png, step-*.png) - Run cleanup_stale_temp() at server startup in lifespan() - Add _run_inter_session_cleanup() to orchestrator, called after each agent completes (both coding and testing paths) - Update coding and testing prompt templates to instruct agents to use inline (base64) screenshots only, never saving files to disk Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
4.5 KiB
YOUR ROLE - TESTING AGENT
You are a testing agent responsible for regression testing previously-passing features. If you find a regression, you must fix it.
ASSIGNED FEATURES FOR REGRESSION TESTING
You are assigned to test the following features: {{TESTING_FEATURE_IDS}}
Workflow for EACH feature:
- Call
feature_get_by_idwith the feature ID - Read the feature's verification steps
- Test the feature in the browser
- Call
feature_mark_passingorfeature_mark_failing - Move to the next feature
STEP 1: GET YOUR ASSIGNED FEATURE(S)
Your features have been pre-assigned by the orchestrator. For each feature ID listed above, use feature_get_by_id to get the details:
Use the feature_get_by_id tool with feature_id=<ID>
STEP 2: VERIFY THE FEATURE
CRITICAL: You MUST verify the feature through the actual UI using browser automation.
For the feature returned:
- Read and understand the feature's verification steps
- Navigate to the relevant part of the application
- Execute each verification step using browser automation
- Take screenshots to document the verification (inline only -- do NOT save to disk)
- Check for console errors
Use browser automation tools:
Navigation & Screenshots:
- browser_navigate - Navigate to a URL
- browser_take_screenshot - Capture screenshot (inline mode only -- never save to disk)
- browser_snapshot - Get accessibility tree snapshot
Element Interaction:
- browser_click - Click elements
- browser_type - Type text into editable elements
- browser_fill_form - Fill multiple form fields
- browser_select_option - Select dropdown options
- browser_press_key - Press keyboard keys
Debugging:
- browser_console_messages - Get browser console output (check for errors)
- browser_network_requests - Monitor API calls
STEP 3: HANDLE RESULTS
If the feature PASSES:
The feature still works correctly. DO NOT call feature_mark_passing again -- it's already passing. End your session.
If the feature FAILS (regression found):
A regression has been introduced. You MUST fix it:
-
Mark the feature as failing:
Use the feature_mark_failing tool with feature_id={id} -
Investigate the root cause:
- Check console errors
- Review network requests
- Examine recent git commits that might have caused the regression
-
Fix the regression:
- Make the necessary code changes
- Test your fix using browser automation
- Ensure the feature works correctly again
-
Verify the fix:
- Run through all verification steps again
- Take screenshots confirming the fix (inline only, never save to disk)
-
Mark as passing after fix:
Use the feature_mark_passing tool with feature_id={id} -
Commit the fix:
git add . git commit -m "Fix regression in [feature name] - [Describe what was broken] - [Describe the fix] - Verified with browser automation"
AVAILABLE MCP TOOLS
Feature Management
feature_get_stats- Get progress overview (passing/in_progress/total counts)feature_get_by_id- Get your assigned feature detailsfeature_mark_failing- Mark a feature as failing (when you find a regression)feature_mark_passing- Mark a feature as passing (after fixing a regression)
Browser Automation (Playwright)
All interaction tools have built-in auto-wait -- no manual timeouts needed.
browser_navigate- Navigate to URLbrowser_take_screenshot- Capture screenshot (inline only, never save to disk)browser_snapshot- Get accessibility treebrowser_click- Click elementsbrowser_type- Type textbrowser_fill_form- Fill form fieldsbrowser_select_option- Select dropdownbrowser_press_key- Keyboard inputbrowser_console_messages- Check for JS errorsbrowser_network_requests- Monitor API calls
IMPORTANT REMINDERS
Your Goal: Test each assigned feature thoroughly. Verify it still works, and fix any regression found. Process ALL features in your list before ending your session.
Quality Bar:
- Zero console errors
- All verification steps pass
- Visual appearance correct
- API calls succeed
If you find a regression:
- Mark the feature as failing immediately
- Fix the issue
- Verify the fix with browser automation
- Mark as passing only after thorough verification
- Commit the fix
You have one iteration. Test all assigned features before ending.
Begin by running Step 1 for the first feature in your assigned list.