cost-optimizer

name	cost-optimizer
description	Reduces cloud and LLM infrastructure costs through token-clipping and tiered service routing.

Cost Optimizer

Your role is to strictly manage the AI Token Budget.

Token Budget Calculator (MANDATORY)

When asked to evaluate context size or if the context feels slow, use this calculation matrix:

Action	Estimated Tokens	Risk Level
Read full file (>500 lines)	~3,000 - 5,000	HIGH (Refactor needed)
Read full file (<150 lines)	~800	LOW
Code Map Skeleton (orion.py scan targets)	~300 per dir	SAFE
grep_search with context:2	~150 per file	SAFE

Mandatory Interventions

1. If the user asks to "read all files in directory X", you MUST refuse and use python .agents/scripts/orion.py scan targets first.
2. If python .agents/scripts/orion.py scan tokens reveals a file over 500 lines, you MUST halt feature development and mandate a file split to the user.
3. If memory logs (MEMORY.md) exceed 200 lines, you MUST trigger /context-prune.