Vercel Efficiency Audit

You are an AI coding agent performing a READ-ONLY audit of this repository to identify Vercel cost/efficiency improvements.

## CONSTRAINTS (non-negotiable)
1. NO MODIFICATIONS to existing files—no edits, patches, formatting, PRs, or auto-fixes
2. Create exactly ONE file: `docs/audits/VERCEL_EFFICIENCY_AUDIT_XX.md` (XX = smallest unused number starting at 01)
3. Audit the ENTIRE repo: app code, API routes, functions, edge/middleware, cron, configs, build pipeline, dependencies

## COST DRIVERS TO INVESTIGATE
- **Runtime**: Slow handlers, cold-start sensitivity, unnecessary per-request work
- **Invocations**: Chatty endpoints, polling, unbounded cron, webhook storms
- **Bandwidth**: Large responses, uncompressed assets, missing cache headers
- **Build**: Heavy bundling, no caching, bloated dependencies
- **Assets**: Oversized images/media, missing optimization/resizing
- **Data access**: N+1 queries, repeated API calls, no memoization/caching
- **Rendering**: Dynamic where static works (no ISR/SSG leverage)
- **External calls**: No timeouts/retries/backoff in hot paths
- **Logging**: Verbose tracing in production hot paths

## WORKFLOW
1. **Inventory**: Map execution surfaces (routes, functions, edge, SSR, cron), build tools, heavy computation/data locations
2. **Analyze**: Find cost/perf opportunities—prioritize caching, reducing work/invocations, shrinking payloads, build optimization
3. **Report**: Write findings using the template below

## REPORT TEMPLATE
```markdown
# Vercel Efficiency Audit (XX)
_Date: YYYY-MM-DD | Repository: <name> | Auditor: AI (read-only)_

## Executive Summary
- Top 5 cost drivers
- Top 10 actions (ranked)
- Quick wins (<1 day) | Medium lifts (1-3 days) | Bigger bets (multi-day)

## Cost Surface Inventory
- Runtime surfaces (API routes, functions, edge, middleware, SSR)
- Background/cron jobs
- High-traffic endpoints (inferred, note uncertainty)
- Build pipeline summary
- Large assets & heavy dependencies

## Findings

### 1. [Title]
| Attribute | Detail |
|-----------|--------|
| Severity | Critical/High/Medium/Low |
| Cost Driver | runtime/invocations/bandwidth/build/other |
| Evidence | File paths, function names, brief description |
| Recommendation | What to change (no code patches) |
| Expected Impact | e.g., "reduce invocations ~30%", "cut payload ~50%" |
| Effort | S/M/L |
| Risks/Tradeoffs | |
| Validation | How to verify (metrics, tests) |

[Repeat for each finding...]

## Cross-Cutting Recommendations
- Caching strategy (HTTP/CDN/app-level)
- External call standards (timeouts, retries, backoff)
- Prod logging guidelines
- Performance budgets (payload size, TTFB, function duration)

## Measurement Plan
- Metrics: TTFB, p95 latency, function duration, invocations, response sizes, cache hit rate, build time
- Instrumentation approach
- Rollout strategy

## Appendix
- Files reviewed
- Assumptions & unknowns
```

## BEFORE SUBMITTING
Verify: ✓ No files modified ✓ Created exactly one new file with correct naming ✓ All recommendations reference specific file paths

You are an AI coding agent performing a READ-ONLY audit of this repository to identify Vercel cost/efficiency improvements.

## CONSTRAINTS (non-negotiable)
1. NO MODIFICATIONS to existing files—no edits, patches, formatting, PRs, or auto-fixes
2. Create exactly ONE file: `docs/audits/VERCEL_EFFICIENCY_AUDIT_XX.md` (XX = smallest unused number starting at 01)
3. Audit the ENTIRE repo: app code, API routes, functions, edge/middleware, cron, configs, build pipeline, dependencies

## COST DRIVERS TO INVESTIGATE
- **Runtime**: Slow handlers, cold-start sensitivity, unnecessary per-request work
- **Invocations**: Chatty endpoints, polling, unbounded cron, webhook storms
- **Bandwidth**: Large responses, uncompressed assets, missing cache headers
- **Build**: Heavy bundling, no caching, bloated dependencies
- **Assets**: Oversized images/media, missing optimization/resizing
- **Data access**: N+1 queries, repeated API calls, no memoization/caching
- **Rendering**: Dynamic where static works (no ISR/SSG leverage)
- **External calls**: No timeouts/retries/backoff in hot paths
- **Logging**: Verbose tracing in production hot paths

## WORKFLOW
1. **Inventory**: Map execution surfaces (routes, functions, edge, SSR, cron), build tools, heavy computation/data locations
2. **Analyze**: Find cost/perf opportunities—prioritize caching, reducing work/invocations, shrinking payloads, build optimization
3. **Report**: Write findings using the template below

## REPORT TEMPLATE
```markdown
# Vercel Efficiency Audit (XX)
_Date: YYYY-MM-DD | Repository: <name> | Auditor: AI (read-only)_

## Executive Summary
- Top 5 cost drivers
- Top 10 actions (ranked)
- Quick wins (<1 day) | Medium lifts (1-3 days) | Bigger bets (multi-day)

## Cost Surface Inventory
- Runtime surfaces (API routes, functions, edge, middleware, SSR)
- Background/cron jobs
- High-traffic endpoints (inferred, note uncertainty)
- Build pipeline summary
- Large assets & heavy dependencies

## Findings

### 1. [Title]
| Attribute | Detail |
|-----------|--------|
| Severity | Critical/High/Medium/Low |
| Cost Driver | runtime/invocations/bandwidth/build/other |
| Evidence | File paths, function names, brief description |
| Recommendation | What to change (no code patches) |
| Expected Impact | e.g., "reduce invocations ~30%", "cut payload ~50%" |
| Effort | S/M/L |
| Risks/Tradeoffs | |
| Validation | How to verify (metrics, tests) |

[Repeat for each finding...]

## Cross-Cutting Recommendations
- Caching strategy (HTTP/CDN/app-level)
- External call standards (timeouts, retries, backoff)
- Prod logging guidelines
- Performance budgets (payload size, TTFB, function duration)

## Measurement Plan
- Metrics: TTFB, p95 latency, function duration, invocations, response sizes, cache hit rate, build time
- Instrumentation approach
- Rollout strategy

## Appendix
- Files reviewed
- Assumptions & unknowns
```

## BEFORE SUBMITTING
Verify: ✓ No files modified ✓ Created exactly one new file with correct naming ✓ All recommendations reference specific file paths