What problem does it solve? Long AI coding sessions can silently accumulate high token costs, and without visibility into spending drivers, users overspend on large file reads, broad searches, and unnecessary subagent spawns. ## Core Features & Use Cases - Cost Awareness: Identifies the main cost drivers such as large file reads, broad grep searches, subagent spawning, and model selection, with concrete optimization tactics for each. - Budget Planning: Provides token and tool-call budget tables by task type (bug fix, small feature, large feature, refactor, code review) with wrap-up thresholds. - Optimization Strategies: Offers ten concrete rules including scoped prompts, right-sized model selection, proactive compaction, and batched tool calls. - Use Case: Midway through a large refactoring session, check current spend against the refactor budget, compact context at the task boundary, and decide whether to continue or start a fresh session. ## Quick Start Ask the assistant to check current session costs and suggest optimizations based on the cost-tracker budgets.
