About this skill

Problem

WorkBuddy local traces store request logs, token usage, model names, errors, and credit signals, but the raw files under ~/.workbuddy are not easy to inspect. Engineers often need answers such as total token usage, expensive models, peak days, and error concentration. This skill focuses on WorkBuddy's own local usage and turns that data into a readable dashboard.

How It Works

The skill uses only the Python standard library, reads local data read-only, and does not modify source files. Key steps:
- Aggregate metrics: daily rollups in local timezone are written to usage-status.json, covering tokens, model mix, cost, errors, and peak days.
- Generate dashboard: it creates workbuddy-usage-status-dashboard-*.html with inlined data and Chart.js, no external network dependencies, and timestamped filenames to avoid overwriting old reports.
- Optional exact credits: when server-side precision is needed, --credit-xlsx can override the relevant date window; by default credits are estimated locally.
- Delivery: after generation, the latest HTML file is opened with present_files for review.

Boundaries

The skill only reads ~/.workbuddy, so it is not for Cursor, VS Code, Trae, Claude, or other product usage, and it is not a general-purpose visualization tool. Metric definitions follow DATA-GUIDE.md: tokens are the authoritative primary metric, while credits are session-level estimates. Some traces may be skipped or fail parsing, so the report may be incomplete and should be disclosed.

Screenshots

Use Cases

During month-end cost review, inspect WorkBuddy local dashboards for tokens, model mix, and peak days.
Investigate WorkBuddy request failures and model distribution in a week to locate error-heavy dates.
When explaining WorkBuddy monthly usage to a team, generate a self-contained HTML report and review it.
When reconciling monthly WorkBuddy credits, import the credit xlsx and inspect the exact date window.

Best For

Engineers maintaining WorkBuddy local usage records who want quick token, error, and peak-day views.
Product or operations leads reconciling costs who need to compare monthly WorkBuddy credit and token definitions.
Developers using WorkBuddy for daily work who want a local HTML dashboard for usage reporting.
Owners auditing agent local data access who need to confirm the tool is read-only and does not upload data.