version: "1.0.0" name: model-usage description: Summarize per-model cost usage from CodexBar JSON for ClawLite workflows. metadata: {"clawlite":{"emoji":"📊"}} script: model_usage

Model usage

Use this skill when the user asks for model-level usage or cost split from CodexBar exports.

What it does

• Reads CodexBar cost --format json payloads
• Summarizes by model for one provider (codex or claude)
• Supports current-model mode (latest day heuristic) and all-model mode

Run manually

From the ClawLite repo root:

python clawlite/skills/model-usage/scripts/model_usage.py --provider codex --mode current
python clawlite/skills/model-usage/scripts/model_usage.py --provider codex --mode all
python clawlite/skills/model-usage/scripts/model_usage.py --provider claude --mode all --format json --pretty

With an input file or stdin:

codexbar cost --provider codex --format json > /tmp/cost.json
python clawlite/skills/model-usage/scripts/model_usage.py --input /tmp/cost.json --mode all
cat /tmp/cost.json | python clawlite/skills/model-usage/scripts/model_usage.py --input - --mode current

Notes

• Script is stdlib-only and can run without ClawLite runtime services.
• CodexBar output usually exposes cost per model; token split by model may be unavailable.