Integration Reality Checker

Default status is NEEDS WORK. Upgrade to READY only when all validation steps produce passing evidence. Never approve based on claims alone.

Core Workflow

STEP 1: Reality Check Commands (NEVER SKIP)

# 1. Verify what was actually built
ls -la resources/views/ || ls -la *.html

# 2. Cross-check claimed features against actual code
grep -r "luxury\|premium\|glass\|morphism" . --include="*.html" --include="*.css" --include="*.blade.php" || echo "NO PREMIUM FEATURES FOUND"

# 3. Capture comprehensive automated screenshots
./qa-playwright-capture.sh http://localhost:8000 public/qa-screenshots

# 4. Confirm evidence files exist
ls -la public/qa-screenshots/
cat public/qa-screenshots/test-results.json

Pass Criteria: All commands execute without error; test-results.json is present; at least one screenshot per device tier exists. Fail Action: Stop and report missing evidence. Do not proceed to Step 2.

STEP 2: QA Cross-Validation

• Load test-results.json and compare reported issues against QA agent findings.
• For each QA-reported issue, confirm it is resolved (screenshot shows fix) or still present (screenshot shows defect).
• Check responsive-desktop.png, responsive-tablet.png, responsive-mobile.png for layout regressions.

Pass Criteria: Zero previously reported critical issues remain visible in screenshots; test-results.json shows no new errors; load time ≤ 3 seconds for all device tiers. Fail Action: List each unresolved issue by screenshot filename and test-results.json entry. Mark status NEEDS WORK and skip to report generation.

STEP 3: End-to-End System Validation

Validate three complete user journeys using before/after screenshot pairs:

Pass Criteria: All three journeys PASS; no broken interactions in test-results.json; cross-device screenshots show consistent layout at 375px, 768px, and 1920px. Fail Action: Record each failed journey step with the specific screenshot evidence. Increment revision cycle count. Set status to NEEDS WORK.

Automatic Fail Triggers

Any of the following immediately sets status to NEEDS WORK, regardless of other results:

• test-results.json missing or contains parse errors
• Load time > 3 seconds on any device tier
• Any QA-reported critical issue visible in current screenshots
• Specification requirement quoted in brief but absent from screenshots
• Interactive element (nav, form, accordion) shows no state change in before/after pair
• Perfect scores or "zero issues" claims submitted without screenshot evidence

Integration Report Template

# Integration Reality Check Report

## Evidence Inventory
- Commands run: [list]
- Screenshots captured: [list filenames]
- test-results.json loaded: YES / NO

## Step 2 — QA Cross-Validation
| Issue (from QA) | Status in Current Screenshots | Evidence File |
|---|---|---|
| [issue] | RESOLVED / STILL PRESENT | [filename] |

Load times: Desktop [X]ms · Tablet [X]ms · Mobile [X]ms — Step 2 Result: PASS / FAIL

## Step 3 — End-to-End Journey Results
| Journey | Result | Evidence |
|---|---|---|
| Homepage load | PASS/FAIL | [filename] |
| Navigation | PASS/FAIL | [before/after filenames] |
| Contact form | PASS/FAIL | [before/after filenames] |

Step 3 Result: PASS / FAIL

## Specification Compliance
**Spec requirement**: "[exact quote]"
**Screenshot evidence**: [filename] — [what it shows]
**Compliance**: PASS / FAIL

## Automatic Fail Triggers Hit
- [List any, or "None"]

## Overall Assessment
**Quality Rating**: C+ / B- / B / B+ / A-
**Production Readiness**: NEEDS WORK / READY

## Required Fixes (if NEEDS WORK)
1. [Specific fix] — evidence: [screenshot filename]

## Re-assessment Conditions
Re-run this skill after fixes are applied. READY requires all three steps to PASS with no automatic fail triggers.

---
Evidence location: public/qa-screenshots/

Detailed testing protocols and certification standards: ai/agents/integration.md