Token Usage Dashboard

Use CodexBar CLI local cost usage to summarize per-model usage for Codex or Claude, including the current (most recent) model or a full model breakdown. Trig...

This package appears to do what it says: it runs the CodexBar CLI (or ingests an exported JSON) and builds a local dashboard. Before installing or running it: 1) Inspect the brew tap (steipete/tap) and the codexbar formula to ensure you trust that source. Adding a third‑party Homebrew tap can install arbitrary code. 2) Be aware the dashboard parses local LLM session logs (via CodexBar) and llmCalls can include raw prompt text — treat outputs and any generated summary files as sensitive. 3) Example alert configs include webhook URLs, but active dispatch is reportedly not implemented; nonetheless, check any alert/config you provide so you don't accidentally post sensitive data to external webhooks. 4) If you need to be extra cautious, run the scripts on exported JSON copies (not on live logs) in a controlled environment and review code before executing.

Type: OpenClaw Skill Name: token-usage-dashboard Version: 1.0.0 The token-usage-dashboard skill bundle is a legitimate tool designed to visualize and manage LLM costs and usage patterns using data from the CodexBar CLI. The core logic in scripts/token_usage_dashboard.py and scripts/model_usage.py focuses on local data processing, including cost attribution, anomaly detection, and multi-tenant access control. The implementation demonstrates security awareness through the use of safe subprocess calls, XSS protection in the generated HTML dashboard, and automated anonymization of sensitive prompt data. No indicators of data exfiltration, malicious persistence, or unauthorized remote execution were found, and the bundle includes comprehensive unit tests and documentation.