Melvynx

Agents Analysis — Development skill for Claude Code

Development community

Effective token price reports for Codex and Claude agent subscriptions.

How to install Agents Analysis

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

What Agents Analysis does

Effective token price reports for Codex and Claude agent subscriptions.

Alternatives in Development

Ccusage — CLI for analyzing Claude Code/Codex usage from local JSONL files 11.8k ★
Dao Code — Open-source TypeScript terminal coding agent for DeepSeek-V4 — builds on DeepSeek's strong price-performance a 1.3k ★
Pool Ohlcv — Fetch price and volume history for a specific Meteora pool 722 ★

README

Agents Analysis

Open source examples and prompts for calculating the effective token price of AI agent subscriptions from local usage logs.

The goal is simple: compare what a user pays for a subscription with the API-equivalent value of the usage they actually consumed.

This repository currently focuses on:

Codex usage from .codex logs.
Claude Code usage from .claude logs.
Effective price per 1M tokens.
Weekly quota extrapolation from the usage percentage shown in the product UI.
Small helper scripts to recompute results faster.
Markdown pricing references for the model baselines used in examples.

API-equivalent value is not the provider's real cost. It is the price a user would roughly pay if the same token mix were reproduced through official public API pricing.

Repository Structure

result/
  /
    codex/
      /
        data.json
        analysis.md
    claude/
      /
        data.json
        analysis.md
CONTRIBUTE.md
README.md

Each contributor should add results under their own GitHub username:

result/alice/codex/2026-05-30T21-14-45Z/data.json
result/alice/codex/2026-05-30T21-14-45Z/analysis.md
result/alice/claude/2026-05-30T21-14-45Z/data.json
result/alice/claude/2026-05-30T21-14-45Z/analysis.md

Each analysis run is stored in its own timestamped folder under the selected tool. This allows multiple analyses during the same weekly quota window without overwriting earlier results. Use the measurement end time in UTC for the folder name, formatted like `YYYY-MM-DDTHH-MM-SSZ`.