Natural language time tracking with smart activity classification, multi-dimensional reports, productivity analytics, and billing.
Documents
Computer Usage Summary
Try itPrivate local app, project, and billable reports
What it does
Summarize what a user did on macOS, Windows, or Linux using local ActivityWatch records, with privacy-preserving live-state and Screen Time fallbacks when records are unavailable. Use for daily activity summaries, app usage time, foreground sessions, AFK time, timelines, and paste-ready spreadsheet tables.
The skill document
Computer Usage Summary
Overview
Use local ActivityWatch data first. This skill supports macOS, Windows, and
Linux wherever the local ActivityWatch API and Python 3.9+ are available. The
intended privacy-balanced setup records
only foreground apps, sanitized window titles, and AFK state. Do not install
browser extensions, enable aw-sync, configure cloud storage, or expose the
local API.
Accurate historical app time and AFK time exist only from the time ActivityWatch starts recording. Do not infer them from process uptime, macOS system logs, browser history, or undocumented databases.
Setup And Verification
Use this section only when ActivityWatch is missing, stopped, or needs repair.
- Inspect the local ActivityWatch application, its startup setting, local processes, and the current user's ActivityWatch data directory. Do not replace a working installation.
- Let the user choose an official installation source. Before launching it or
enabling automatic startup, verify the publisher and operating-system
security status. On macOS, use
codesign --verify --deep --strictandspctl --assess --type execute. Stop if validation fails unless the user explicitly decides otherwise. - Enable only the window and AFK watchers. Let macOS request Accessibility permission through its normal interface; never bypass or change privacy permissions programmatically.
- On Windows, run
python -m pip install tzdataonce so Python can resolve IANA time zones such asAsia/Singaporeconsistently. - Confirm the local
http://127.0.0.1:5600server hascurrentwindowandafkstatusbuckets containing events. The default AFK threshold is 180 seconds.
No-ActivityWatch Fallback
When the app, local API, buckets, or requested-range events are unavailable:
- Mark active time, AFK time, sessions, and launch counts as
unavailable, not zero. - A current frontmost-app or running-app snapshot is only evidence of the present, not the requested range.
- Use Screen Time only when the user makes its visible totals available; label
it
Screen Time reportedand do not read private databases. - Inspect browser history only for an explicit website-level request. Inspect recent documents or terminal history only after explicit, source-specific permission. Treat all such results as metadata clues, not durations.
Workflow
- Establish the local date range and time zone.
- Run
scripts/activitywatch_summary.py --date today, use--period weekor--period montharound an anchor date, or use inclusive--start YYYY-MM-DD --end YYYY-MM-DDdates. Add--timezone Area/Cityfor reproducible local timestamps. Keep--api-urlon the loopback interface. - Use
active_secondsandafk_secondsonly when ActivityWatch confirms them. Describeforeground_sessionsas observed foreground intervals, not process launches. - Use sanitized app names and window titles to reconstruct the timeline. A
title is
confirmedonly when it directly supports the described activity. - When the user needs project/client or billable reporting, offer an optional
local JSON
--rulesfile. Match only sanitized titles. Explain that the first matching rule wins and never create or upload rules without consent. - For spreadsheet output, run one table per command:
--format markdown --table appsfor chat.--format tsv --table appsto copy directly into spreadsheet software.--format tsv --table timelinefor chronological rows.--format csv --table apps --csv-bom --output report.csvfor an Excel-friendly file, only when the user asks to save one.--report client-timesheetfor project/client billable rows.--report weekly-reviewfor a compact review table.--report app-trendfor local daily trend rows.
- When ActivityWatch is unavailable, return its structured reason. Markdown,
TSV, and CSV outputs provide
status,reason, andsourcecolumns.
Default Conversational Output
For conversational summaries, always use the following four sections, in this order, unless the user explicitly asks for a simpler or shorter output:
Today at a glance- State the local date and time zone.
- Report confirmed active time, away time, and evidence coverage.
- In fallback mode, state that unavailable values are
unavailable; do not estimate them.
What you did- Give a chronological list of meaningful activity windows.
- Include the time window, activity, supporting app or sanitized title,
confidence, and evidence label (
confirmed ActivityWatch,Screen Time reported,live snapshot, ormetadata clue).
App table- Include app name, observed foreground sessions, active time, first seen, last active, and source note when ActivityWatch data is available.
- If data is unavailable, keep this section and show the structured reason
with
unavailablemetrics rather than omitting the table.
Caveats- State the relevant privacy and evidence limits, including that foreground sessions are observed intervals rather than process-launch counts and that tracking starts only when ActivityWatch starts.
Keep these sections even when one contains little or no data. A request such as "简单一点", "简要输出", "只要结论", or "只给表格" is an explicit request for a reduced format; then omit or compress only the sections the user asks to remove. Do not infer a reduced format merely because the request is short.
For a requested export (TSV, CSV, or a saved report), follow the export
format the user requested instead of wrapping it in these conversational
headings. If the user asks for both a summary and an export, provide the four
conversational sections first and then the requested export.
Privacy Rules
- Treat window titles as sensitive. Remove URLs, truncate long titles, and never include complete browser URLs.
- Keep default sanitized titles for a useful timeline; use
--hide-titlesif the user wants a title-free export. - Use the smallest useful amount of title context; application and summary tables do not need titles.
- Escape spreadsheet cells that could be interpreted as formulas.
- Keep reports local to the current conversation. Do not export, upload, or share activity data without explicit user approval.
Related skills
Productivity tracker and daily review assistant. Input activity logs, time notes, goals, or a daily schedule; output time categories, bottlenecks, focus leak...
Screen activity tracking. Trigger: track screen activity, start/stop tracking, daily summary, search history, 开始追踪屏幕, 屏幕活动总结, 整理今天的操作. Uses cron tool for tra...
Minimal screen activity tracking. Trigger: track screen, start/stop tracking, daily summary, search history, 开始追踪屏幕, 屏幕活动总结, 整理今天的操作. Uses cron tool for trac...
Use the Health Data AI Analyzer Mac app read-only localhost API on macOS to generate a concise Apple Health daily brief and 3 practical suggestions. Only run...
Ingest Apple Health Auto Export JSON (HealthMetrics + Workouts) into a local DuckDB database and render offline HTML dashboards plus a Markdown daily summary...