Installs into .claude/skills of the current project.
Are you the author of Complexity Tier Model Mapping?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/vamseeachanta-complexity-tier-model-mapping)
---
name: agent-usage-optimizer-complexity-tier-model-mapping
description: "Sub-skill of agent-usage-optimizer: Complexity Tier \u2192 Model Mapping."
version: 1.0.0
category: ai
type: reference
scripts_exempt: true
---
# Complexity Tier → Model Mapping
## Complexity Tier → Model Mapping
Use this alongside Route mapping when task nature is clear:
| Complexity tier | Keywords | Recommended model |
|----------------|----------|------------------|
| `routine` | format, rename, config, scaffold, update-doc, copy | Codex Haiku |
| `standard` | implement, review, test, fix, migrate, document | Codex Sonnet |
| `complex` | architecture, design, cross-repo, security, compound | Codex Opus |
**Key insight**: the gap between theoretical and observed exposure is an implementation gap —
not a capability gap. Routing routine tasks to cheaper models closes this gap for our workflow.
See WRK-5002 to automate tier detection in `task_classifier.sh`.
---
*Use this skill before any multi-item work session or when quota is a concern.*
*Related: `ai/optimization/model-selection`, `ai/optimization/usage-optimization`*
