This skill allows an agent to extract and summarize detailed cost and usage data from CodexBar's local logs. It can provide a snapshot of the most recently used model (current model) or a comprehensive breakdown of all models used across different providers.
Use this skill when the user asks for model-level usage statistics, current session costs, or a scriptable summary of spending for AI providers like Codex or Claude.
model_usage.py) that processes cost JSON data.Designed for agents integrated with the Vela/CodexBar ecosystem (Codex, Claude).
Well-crafted skill for summarizing CodexBar per-model cost data. Main script failed due to missing codexbar CLI (macOS-only brew cask), but test suite passed cleanly. Code is well-structured with proper argparse, type hints, dataclasses, and error handling. No security concerns — subprocess calls use list form, no network exposure, no credentials.
codexbarClean skill with good code quality. Tests are included and pass. Niche audience — only CodexBar users on macOS. Architecture follows skill spec well with frontmatter, scripts/, and references/.