# Performance Optimization

> Agent and model performance (Haiku/Sonnet/Opus selection, context window, build troubleshooting). For code performance use vercel-react-best-practices and granular performance skills.

- Skill: `diegosouzapw/performance-optimization-3` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add diegosouzapw/performance-optimization-3`
- Raw SKILL.md: https://api.skillmd.com/api/skills/diegosouzapw/performance-optimization-3/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Web & Frontend
- Author: diegosouzapw (https://skillmd.com/u/diegosouzapw)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/diegosouzapw/performance-optimization-3

---


# Performance Optimization (Agent & Model)

This skill covers **AI agent and model** performance (which model to use, context management, build fixes). For **application/code** performance (React, Next.js, bundles, rendering) use **vercel-react-best-practices** and the granular performance skills (e.g. cache-property-access-in-loops, dynamic-imports-for-heavy-components).

## Model Selection Strategy

**Haiku 4.5** (90% of Sonnet capability, 3x cost savings):

- Lightweight agents with frequent invocation
- Pair programming and code generation
- Worker agents in multi-agent systems

**Sonnet 4.5** (Best coding model):

- Main development work
- Orchestrating multi-agent workflows
- Complex coding tasks

**Opus 4.5** (Deepest reasoning):

- Complex architectural decisions
- Maximum reasoning requirements
- Research and analysis tasks

## Context Window Management

Avoid last 20% of context window for:

- Large-scale refactoring
- Feature implementation spanning multiple files
- Debugging complex interactions

Lower context sensitivity tasks:

- Single-file edits
- Independent utility creation
- Documentation updates
- Simple bug fixes

## Ultrathink + Plan Mode

For complex tasks requiring deep reasoning:

1. Use `ultrathink` for enhanced thinking
2. Enable **Plan Mode** for structured approach
3. "Rev the engine" with multiple critique rounds
4. Use split role sub-agents for diverse analysis

## Build Troubleshooting

If build fails:

1. Use **build-error-resolver** agent
2. Analyze error messages
3. Fix incrementally
4. Verify after each fix