speedrun

# Speedrun Skill

Comprehensive optimization hunting that dispatches parallel agents to find performance bottlenecks across all layers, then applies fixes with mandatory verification.

## The Iron Law

```
NO OPTIMIZATION CLAIM WITHOUT MEASUREMENT.
```

Every optimization must have before/after metrics. Unverified "improvements" are worthless.

## Quick Reference

| Hunter | Archetype | Focus | Color |
|--------|-----------|-------|-------|
| bundle-hunter | OBSESSIVE | File sizes, tree-shaking, code splitting | cyan |
| complexity-hunter | PEDANTIC | Cyclomatic complexity, nesting, function length | blue |
| dead-code-hunter | SUSPICIOUS | Unused exports, orphan files, unreachable paths | yellow |
| dependency-hunter | WARY | Bloat, duplicates, outdated, vulnerabilities | magenta |
| build-hunter | IMPATIENT | Compilation time, caching, parallelization | red |
| algorithm-hunter | ANALYTICAL | Big O, data structures, repeated computation | green |
| query-hunter | SKEPTICAL | N+1, missing indexes, slow queries | green |
| web-vitals-hunter | EMPATHETIC | LCP, CLS, INP, TTFB | yellow |
| memory-hunter | PARANOID | Leaks, allocation patterns, GC pressure | red |
| image-hunter | METICULOUS | Formats, compression, lazy loading | cyan |

## Workflow

1. **Measure baseline** - Gather current metrics (bundle size, build time, complexity scores)
2. **Analyze codebase** - Detect tech stack, structure, applicable hunters
3. **Inject context** - Pass baseline metrics and stack info to all agents
4. **Dispatch ALL applicable hunters in PARALLEL**
5. **Collect findings** - Aggregate with ROI scoring (Impact / Effort)
6. **Write SPEEDRUN.md** - Prioritized optimization report
7. **Apply fixes** - Safe fixes auto-apply, risky fixes require approval
8. **Verify improvements** - Re-measure and report deltas

## Fix Tiers

### Auto-Apply (Safe)
- Remove unused imports/exports
- Delete dead code files
- Optimize image formats
- Add missing lazy loading
- Remove unused dependencies