Dev Frontend Performance Testing (Gate 6)

Overview

Ensure all frontend pages meet Core Web Vitals thresholds, achieve Lighthouse Performance > 90, maintain bundle size within budget, and minimize client component usage.

Core principle: Performance is a feature. Users on slow devices and connections deserve a fast experience. Performance budgets are enforced, not suggested.

<block_condition>

• LCP > 2.5s on any page = FAIL
• CLS > 0.1 on any page = FAIL
• INP > 200ms on any page = FAIL
• Lighthouse Performance < 90 = FAIL
• Bundle size increase > 10% without justification = FAIL </block_condition>

CRITICAL: Role Clarification

This skill ORCHESTRATES. Frontend QA Analyst Agent (performance mode) EXECUTES.

Who	Responsibility
This Skill	Gather requirements, dispatch agent, track iterations
QA Analyst Frontend Agent	Run Lighthouse, measure CWV, analyze bundles, audit components

---

Standards Reference

Standards Source (Cache-First Pattern): This sub-skill reads standards from state.cached_standards populated by dev-cycle Step 1.5. If invoked outside a cycle (standalone), it falls back to direct WebFetch with a warning. See shared-patterns/standards-cache-protocol.md for protocol details.

MANDATORY: Load testing-performance.md standards using the cache-first pattern below.

Required URL: https://raw.githubusercontent.com/LerianStudio/ring/main/dev-team/docs/standards/frontend/testing-performance.md

For the required standards URL above:
  IF state.cached_standards[url] exists:
    → Read content from state.cached_standards[url].content
    → Log: "Using cached standard: {url} (fetched {state.cached_standards[url].fetched_at})"
  ELSE:
    → WebFetch url (fallback — should not happen if orchestrator ran Step 1.5)
    → Log warning: "Standard {url} was not pre-cached; fetched inline"

<fetch_required> https://raw.githubusercontent.com/LerianStudio/ring/main/dev-team/docs/standards/frontend/testing-performance.md </fetch_required>

---

Step 1: Validate Input

REQUIRED INPUT:
- unit_id: [TASK id — this gate runs at task cadence, aggregating all subtasks]
- implementation_files: [union of changed files across all subtasks of the task]
- gate0_handoffs: [array of per-subtask implementation handoffs, one per subtask]

OPTIONAL INPUT:
- performance_baseline: [previous metrics for comparison]
- gate5_handoff: [full Gate 5 output]

if any REQUIRED input is missing:
  → STOP and report: "Missing required input: [field]"

if gate0_handoffs is not an array:
  → STOP and report: "gate0_handoffs must be an array of per-subtask handoffs"

Step 2: Dispatch Frontend QA Analyst Agent (Performance Mode)

Task tool:
  subagent_type: "ring:qa-analyst-frontend"
  prompt: |
    **MODE:** PERFORMANCE TESTING (Gate 6)

    **Standards:** Load testing-performance.md

    **Input:**
    - Task ID: {unit_id} (task-level — aggregates all subtasks)
    - Implementation Files (union across all subtasks): {implementation_files}
    - Per-Subtask Gate 0 Handoffs: {gate0_handoffs}
    - Baseline: {performance_baseline or "N/A"}

    **Scope:** Validate performance for the task (all subtasks aggregated), not a single subtask.

    **Requirements:**
    1. Measure Core Web Vitals (LCP, CLS, INP) on all affected pages
    2. Run Lighthouse audit (Performance score > 90)
    3. Analyze bundle size change vs baseline
    4. Audit 'use client' usage (should be < 40% of components)
    5. Detect performance anti-patterns (bare <img>, useEffect for fetching, etc.)
    6. Verify sindarian-ui imports are tree-shakeable

    **Output Sections Required:**
    - ## Performance Testing Summary
    - ## Core Web Vitals Report
    - ## Handoff to Next Gate

Step 3: Evaluate Results

Parse agent output:

if "Status: PASS" in output:
  → Gate 6 PASSED
  → Return success with metrics

if "Status: FAIL" in output:
  → Dispatch fix to implementation agent (ring:frontend-engineer)
  → Re-run performance tests (max 3 iterations)
  → If still failing: ESCALATE to user

Step 4: Generate Output

## Performance Testing Summary
**Status:** {PASS|FAIL}
**LCP:** {value}s (< 2.5s)
**CLS:** {value} (< 0.1)
**INP:** {value}ms (< 200ms)
**Lighthouse:** {score} (> 90)
**Bundle Change:** {+X%} (< 10%)

## Core Web Vitals Report
| Page | LCP | CLS | INP | Status |
|------|-----|-----|-----|--------|
| {page} | {value} | {value} | {value} | {PASS|FAIL} |

## Bundle Analysis
| Metric | Current | Baseline | Change | Status |
|--------|---------|----------|--------|--------|
| Total JS (gzipped) | {size} | {size} | {change}% | {PASS|FAIL} |

## Server Component Audit
| Metric | Value |
|--------|-------|
| Total .tsx files | {count} |
| Client components | {count} |
| Client ratio | {percent}% (< 40%) |

## Anti-Pattern Detection
| Pattern | Occurrences | Status |
|---------|-------------|--------|
| Bare <img> | {count} | {PASS|FAIL} |
| useEffect for fetching | {count} | {PASS|FAIL} |
| Wildcard sindarian imports | {count} | {PASS|FAIL} |

## Handoff to Next Gate
- Ready for Gate 7 (Code Review): {YES|NO}
- Iterations: {count}

---

Severity Calibration

Severity	Criteria	Examples
CRITICAL	Core Web Vitals fail, page unusable	LCP > 4s, CLS > 0.25, INP > 500ms
HIGH	Threshold violations, major regressions	Lighthouse < 90, bundle +20%, LCP > 2.5s
MEDIUM	Minor threshold concerns, optimization opportunities	Client ratio > 40%, bare img tags
LOW	Best practices, minor optimizations	Code splitting suggestions, cache improvements

Report all severities. CRITICAL = immediate fix (UX broken). HIGH = fix before gate pass. MEDIUM = fix in iteration. LOW = document.

---

Anti-Rationalization Table

See shared-patterns/shared-anti-rationalization.md for universal anti-rationalizations. Gate-specific:

Rationalization	Why It's WRONG	Required Action
"Works fine on my machine"	Your machine ≠ user's device. Measure objectively.	Run Lighthouse
"We'll optimize later"	Performance debt compounds. Fix during development.	Meet thresholds now
"Bundle size doesn't matter"	Mobile 3G users exist. Every KB matters.	Stay within budget
"Everything needs use client"	Server components reduce JS. Audit first.	Minimize client components
"next/image is too complex"	next/image gives free optimization. Always use it.	Use next/image
"Lighthouse 85 is close enough"	90 is the threshold. 85 = FAIL.	Optimize to 90+

---