返回 Skill 列表
extension
分类: AI Agent 能力无需 API Key

what-did-i-do

被动屏幕活动追踪器。当用户询问他们今天做了什么、想要追踪他们的屏幕活动或需要分析他们的生产力模式时,使用此技能。

person作者: jakexiaohubgithub

what-did-i-do

Passive screen tracker that captures screenshots every 5 minutes, then uses Gemini Vision to analyze what you actually did.

Commands

Use Bash to run these commands:

# Start recording screenshots
~/.config/opencode/skill/what-did-i-do/scripts/watcher.sh start

# Stop recording
~/.config/opencode/skill/what-did-i-do/scripts/watcher.sh stop

# Check status
~/.config/opencode/skill/what-did-i-do/scripts/watcher.sh status

# Analyze pending screenshots (sends to Gemini)
~/.config/opencode/skill/what-did-i-do/scripts/watcher.sh analyze

Reading Daily Reports

Daily activity logs are stored in ~/.config/opencode/skill/what-did-i-do/daily/YYYY-MM-DD.md

Use Read to view today's report:

~/.config/opencode/skill/what-did-i-do/daily/2026-02-01.md

Report Format

Each daily file contains:

  • Time blocks with activity tables
  • Focus categories: Deep work / Light work / Break / Distraction
  • Summary with minutes per category

Common Use Cases

| User Says | Action | |-----------|--------| | "What did I do today?" | Read today's daily file, summarize | | "Start tracking my screen" | Run watcher.sh start | | "Stop tracking" | Run watcher.sh stop | | "Analyze my day" | Run watcher.sh analyze, then read the output | | "How productive was I?" | Read daily file, calculate deep work % | | "What did I do this week?" | Read last 7 daily files, summarize patterns |

Setup Required

User must set their Gemini API key in scripts/analyze.py (line 12). Get one free at: https://aistudio.google.com/apikey