gaia-research

Skill Cost — AI skill for Claude Code

AI community

Multi-harness token-usage cost reporter for AI agent sessions (pi, Claude Code, Codex, opencode).

How to install Skill Cost

This entry records only its repository, not the path inside it, so there is no exact command to give. Open gaia-research/skill-cost and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Skill Cost does

Multi-harness token-usage cost reporter for AI agent sessions (pi, Claude Code, Codex, opencode). Reads JSONL logs, prices from LiteLLM catalog, auto-refresh built-in. One file, zero deps.

Alternatives in AI

Codeburn — Free, local tool to track AI coding token usage and cost across 37 tools and agents (Claude Code, Cursor, Code 9.7k ★
Manifest — Real-time cost observability for OpenClaw agents — track tokens, costs, messages, and model usage 4.1k ★
Surface — Show ranked attack surface for a target from its recon manifest + hunt memory 3.8k ★

README

cost

**Token-usage & USD cost report for AI agent sessions — pi, Claude Code, Codex, opencode, Hermes Agent**

[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](./LICENSE) [![Powered by Gaia](https://img.shields.io/badge/powered%20by-Gaia%20skill--tree-6b46c1)](https://gaiaskilltree.com) [![Prices by LiteLLM](https://img.shields.io/badge/prices%20by-BerriAI%2Flitellm-0ea5e9)](https://github.com/BerriAI/litellm/blob/main/model_prices_and_context_window.json) [![Python 3.8+](https://img.shields.io/badge/python-3.8+-3776ab.svg)](https://www.python.org) [![Zero deps](https://img.shields.io/badge/deps-stdlib%20only-success)](#requirements) [![Prices auto-refresh](https://github.com/gaia-research/skill-cost/actions/workflows/refresh-prices.yml/badge.svg)](./.github/workflows/refresh-prices.yml)

**Prices sourced from [BerriAI/litellm](https://github.com/BerriAI/litellm) (MIT).** **[Powered by Gaia](https://gaiaskilltree.com) — an open registry of AI agent skills.**

**How much did that agent session actually cost?**

Your harness persists token counts in JSONL logs or a local session database. It often does *not* record dollars — pi records `cost: 0`, self-hosted proxies rarely price anything, and the "cost" number many CLIs surface is stale by a version. `cost` reads the persisted token counts and prices them against [LiteLLM's canonical catalog](https://github.com/BerriAI/litellm/blob/main/model_prices_and_context_window.json), producing a reproducible public-rate estimate. Subscription-included or negotiated billing may differ from this estimate.

No pip. No npm. No config. **One Python file, one JSON price sheet, prices auto-refresh every 7 days.**

Why this exists

I kept asking my agent, at the end of long orchestration sessions, "how much did that cost?" It would say things like `~$2` with high confidence. I believed it for a while.

Then I checked the raw token counts. One session it had estimated at $2 was closer to $8 — 4× off