deepwork-tracker

$npx mdskill add aAAaqwq/AGI-Super-Team/deepwork-tracker

Start, stop, and visualize deep work sessions locally.

  • Manages local timers and logs user focus duration.
  • Depends on SQLite storage and macOS Clock integration.
  • Executes commands based on natural language intent.
  • Outputs text reports or Telegram-ready heatmaps.

SKILL.md

.github/skills/deepwork-trackerView on GitHub ↗
---
name: deepwork-tracker
description: Track deep work sessions locally (start/stop/status) and generate a GitHub-contribution-graph style minutes-per-day heatmap for sharing (e.g., via Telegram). Use when the user says things like “start deep work”, “stop deep work”, “am I in a session?”, “show my deep work graph”, or asks to review deep work history.
---

# Deepwork Tracker

Use the local deepwork app (SQLite-backed) at `~/clawd/deepwork/deepwork.js`.

## Bootstrap (if the script is missing)

If `~/clawd/deepwork/deepwork.js` does not exist, bootstrap it from the public repo:

```bash
mkdir -p ~/clawd
cd ~/clawd

# Clone if missing
[ -d ~/clawd/deepwork-tracker/.git ] || git clone https://github.com/adunne09/deepwork-tracker.git ~/clawd/deepwork-tracker

# Ensure expected runtime path exists
mkdir -p ~/clawd/deepwork
cp -f ~/clawd/deepwork-tracker/app/deepwork.js ~/clawd/deepwork/deepwork.js
chmod +x ~/clawd/deepwork/deepwork.js
```

(Do not fail the user request if clone/copy fails—still attempt other steps and report what’s missing.)

## Commands

Run via exec:

- Start a session (also starts a macOS Clock timer; default target 60m):
  - `~/clawd/deepwork/deepwork.js start --target-min 60`
- Stop a session:
  - `~/clawd/deepwork/deepwork.js stop`
- Check status:
  - `~/clawd/deepwork/deepwork.js status`
- Generate a report:
  - Last 7 days (default): `~/clawd/deepwork/deepwork.js report --days 7 --format text`
  - Telegram-ready last 7 days: `~/clawd/deepwork/deepwork.js report --days 7 --format telegram`
  - Heatmap (optional): `~/clawd/deepwork/deepwork.js report --mode heatmap --weeks 52 --format telegram`

## Chat workflows

### Start deep work
1) Run `~/clawd/deepwork/deepwork.js start --target-min 60` (or another target if the user specifies it).
2) This should also start a macOS Clock timer for the target duration (best-effort; may require Accessibility permissions).
3) Reply with the confirmation line.

### Stop deep work
1) Run `~/clawd/deepwork/deepwork.js stop`.
2) Reply with duration.

### Show deep work graph
1) Run `~/clawd/deepwork/deepwork.js report --days 7 --format telegram`.
2) **Always send** the output to Alex on Telegram (id `8551040296`) using the `message` tool with a Markdown monospace code block.
3) Optionally acknowledge in the current chat that it was sent.

If the user wants different ranges, support `--days 7|14|30|60`.
(Heatmap is still available via `--mode heatmap --weeks ...` when explicitly requested.)

More from aAAaqwq/AGI-Super-Team

SkillDescription
a-fund-monitor监控 A 股基金实时估值与盘后净值,自动判断交易日并生成提醒或分析。
add-leadAdd company/person/relationship to CRM
ads-agentAI-агент для управления Facebook рекламой. Вызывай для анализа, оптимизации, создания кампаний и отчётов.
afrexai-compliance-auditRun internal compliance audits against major governance and security
afrexai-personal-financeComplete personal finance system — budgeting, debt payoff, investing, tax optimization, net worth tracking, and financial independence planning. Use when managing money, building wealth, paying off debt, planning retirement, or optimizing taxes. Zero dependencies.
after-salesUse when managing post-purchase experience, building customer loyalty, or increasing repeat purchases
agent-contactsAI agent contacts — add, list, remove MCP contacts. Use when someone gives an agent URL, or when you need to view/remove contacts.
agent-model-switcher批量查看和切换子 agent 的模型配置,用于统一调整多 agent 的 provider/model 设置。
agent-patternsFormat SPAWN REQUEST messages to launch parallel agents, generate structured agent status reports, and define communication protocols within the sprint system. Use when the user needs to coordinate multiple agents, format spawn requests, produce agent reports, or establish inter-agent communication patterns.
agent-task-confirmUse when confirming whether a dispatched agent task was actually received, activated, and progressing after sessions_send or other task handoff actions.