costea

Cost estimation before running a task. Scans session history from OpenClaw, Claude Code, and Codex CLI. Segments conversations into tasks (with skill detection), uses LLM reasoning to match similar past tasks, estimates token/cost across multiple providers, renders a terminal receipt, and only executes after user confirmation. Triggers on: 'costea', 'estimate cost', 'how much will this cost'.

Historical token consumption report across OpenClaw, Claude Code, and Codex CLI. Rebuilds task index from all session logs and produces a structured summary: per-source breakdown, per-skill aggregation, per-model cost, tool usage patterns, and reasoning vs tool-invocation analysis. Use when: 'token report', 'usage summary', 'how much have I spent', 'token history', 'cost report', 'show usage'.