This file contains the prompts used/reconstructed for the video:
Top 10 OpenClaw Automations I Use Regularly on Linux
Note: Some prompts are exact or near-exact from the workflow notes. Others are polished reconstructions based on the final video script and demo behavior, so they are safer to copy, clearer to run, and easier for viewers to adapt.
- Clean Messy Downloads Folder
- Screenshot Organizer
- Voice Note to Transcript to Tasks
- Morning Briefing
- Calendar Planner
- Meeting Prep + Post-Meeting Notes
- Linux Research Bot
- Second Brain
- Background Autonomous Linux Assistant
- Shell Script Creator
Clean and organize my Downloads folder right now.
Target folder:
~/Downloads
Rules:
1. Work only inside ~/Downloads. Do not touch files outside this folder.
2. Do not delete anything.
3. Do not ask for confirmation before moving files.
4. Create these folders only if needed:
- Documents
- Images
- Audio
- Video
- Archives
- Code
- Installers
- Review
5. Move files only when classification is high-confidence.
6. If a file is ambiguous, move it to Review.
7. Skip incomplete or temporary downloads:
- .part
- .crdownload
- .tmp
- .download
8. Put ISO files, AppImages, .deb files, installers, and unusual binaries in Installers.
9. Never overwrite existing files. If a name conflict happens, keep both files by safely renaming the moved file.
10. After finishing, write a log of all actions to:
~/Downloads/_cleanup-log.md
When done, give me a short summary showing:
- folders created
- number of files moved into each folder
- files skipped
- location of the cleanup log
Turn this Downloads cleanup into a recurring OpenClaw automation.
Workflow name:
Downloads Tidy
Use the same rules as before:
Target folder:
~/Downloads
Rules:
1. Work only inside ~/Downloads.
2. Do not touch anything outside this folder.
3. Do not delete anything.
4. Create these folders only if needed:
- Documents
- Images
- Audio
- Video
- Archives
- Code
- Installers
- Review
5. Move files only when classification is high-confidence.
6. Put ambiguous files in Review.
7. Skip incomplete or temporary downloads:
- .part
- .crdownload
- .tmp
- .download
8. Put ISO files, AppImages, .deb files, installers, and unusual binaries in Installers.
9. Never overwrite existing files. If a conflict happens, preserve both files with a safe renamed filename.
10. Write a cleanup log to:
~/Downloads/_cleanup-log.md
Automation requirements:
1. Create or update the appropriate OpenClaw workspace instruction file so this behavior persists.
2. Create a reliable local script for this workflow if that is the safest and most repeatable approach.
3. Schedule it to run every day at 8:00 PM.
4. Use timezone:
Asia/Kolkata
5. Use an isolated cron/session mode if supported.
6. Run silently in the background unless there is an error.
7. After setting it up, show:
- the standing instruction added
- the script path if a script was created
- the cron schedule
- whether the job is enabled
- a manual command I can run to trigger it immediately
Organize my screenshots folder using visual understanding.
Target folder:
~/Pictures/Screenshots
Task:
1. Examine every screenshot visually.
2. Infer what is visible in each screenshot:
- app or website
- main topic
- page/context
- important visible text
3. Rename each screenshot using a clear descriptive filename.
4. Organize the screenshots into sensible subfolders.
Filename format:
YYYY-MM-DD_HH-MM-SS__source__topic.ext
Rules:
1. Work only inside ~/Pictures/Screenshots.
2. Do not delete anything.
3. Do not overwrite existing files.
4. Preserve the original file extension.
5. Use lowercase words.
6. Use hyphens instead of spaces.
7. Keep filenames short but meaningful.
8. If the screenshot content is unclear, put it in a folder named Review.
9. Create only the folders that are needed.
10. After finishing, summarize:
- total screenshots processed
- folders created
- files renamed
- files moved
- unclear files, if any
Do the work now.
Check whether the Whisper transcription skill is enabled and ready.
If any dependencies are missing, do not try to guess silently.
Tell me exactly what needs to be installed.
If system packages are required, give me the terminal commands.
Do not ask me to paste my password into chat.
After the dependencies are ready, configure Whisper with a small fast model suitable for CPU transcription.
Prefer the tiny model unless there is a clear reason not to.
Then help me test it with a short voice note.
Save this to MEMORY.md:
Whenever I send a voice note that starts with “remind me”, follow this workflow:
1. Transcribe the voice note.
2. Extract all actionable tasks from the transcript.
3. Append those tasks as a checklist to:
~/todo.md
4. If the message includes a specific time or date, schedule a cron reminder for that time.
5. Reply back with:
- the transcription
- the extracted tasks
- whether a reminder was scheduled
6. When I later say that a task is done, remove that completed task from todo.md.
7. Do not invent tasks that were not mentioned.
8. If the time is unclear, ask one short clarification question.
9. Keep replies short and practical.
Remind me — finish the thumbnail, follow up with the sponsor, and check the Ubuntu 26.04 LTS release notes. Remind me at 6 PM today.
Generate today's morning briefing.
Rules:
1. Keep the whole briefing concise and highly scannable.
2. Maximum 8 bullet points total.
3. No fluff.
4. No long explanations.
5. If a section has no reliable update, say “No reliable update”.
6. End with:
Today's focus: <one line summary>
Include exactly these sections:
1. Bengaluru weather
- current temperature or today's forecast
- high/low if available
2. Nifty 50
- latest level
- percentage move
- tone: positive, negative, or flat
3. Market pulse
- one short sentence on market outlook based on reliable live cues
4. Linux news
- 2 or 3 short Linux/open-source headlines only
Set up a cron job for this morning briefing.
Name:
Morning briefing
Schedule:
Every day at 8:00 AM
Timezone:
Asia/Kolkata
Mode:
Use an isolated cron/session mode if supported.
Task:
Generate the same morning briefing every day using the rules below:
1. Keep it concise and highly scannable.
2. Maximum 8 bullet points total.
3. Include Bengaluru weather.
4. Include Nifty 50 level and tone.
5. Include one short market pulse.
6. Include 2 or 3 Linux/open-source headlines.
7. End with:
Today's focus: <one line summary>
Delivery:
If WhatsApp announcement is configured, send the briefing to me on WhatsApp.
If not configured, create the cron job first and tell me what extra step is needed for WhatsApp delivery.
After creating it, show:
- job name
- cron expression
- timezone
- session mode
- delivery mode
- whether the job is enabled
openclaw cron add \
--name "Morning briefing" \
--cron "0 8 * * *" \
--tz "Asia/Kolkata" \
--session isolated \
--message "Generate today's morning briefing. Include Bengaluru weather, Nifty 50 level and tone, one short market pulse, and 2 or 3 Linux/open-source headlines. Keep it under 8 bullet points and end with Today's focus: <one line summary>." \
--announce \
--channel whatsapp \
--to "+91XXXXXXXXXX"Analyze my calendar for Monday, April 13, 2026.
Do not change anything.
I want a practical daily planning briefing.
Include:
1. All events in chronological order.
2. First event of the day.
3. Last event of the day.
4. Biggest free blocks.
5. Awkward gaps that are too short for deep work.
6. Back-to-back meetings.
7. A practical recommendation for how I should structure the day.
Rules:
1. Read my local GNOME Calendar.
2. Do not create, edit, move, or delete events.
3. Be concise but useful.
4. If calendar access fails, tell me exactly what failed.
Schedule a 45-minute event called “Thumbnail Review” on Tuesday, April 14, 2026.
Rules:
1. Schedule it after 2:00 PM.
2. Prefer the earliest sensible free slot after 2:00 PM.
3. Do not overlap existing calendar events.
4. Leave a reasonable buffer if possible.
5. Before making changes, back up the calendar data.
6. Add the event to my local GNOME Calendar.
7. After creating it, tell me:
- what time you chose
- why you chose that slot
- which events you avoided
- where the calendar backup was saved
On Thursday, April 16, 2026, extend my existing “Deep Work Block” so that it ends at 4:00 PM instead of 3:30 PM.
Rules:
1. Keep the start time of the Deep Work Block the same.
2. If another meeting conflicts with the extension, move that meeting to the next sensible 30-minute free slot later the same day.
3. Do not move any other event unless absolutely necessary.
4. Do not delete anything.
5. Back up the calendar data before making changes.
6. After making changes, summarize:
- what event was extended
- what conflict was found
- what event was moved
- the new time for the moved event
- backup location
Add these standing instructions to my OpenClaw workspace:
When I ask for meeting preparation, create a concise, structured, practical meeting brief.
Do not invent facts, decisions, deadlines, attendees, or previous status.
If something is unknown, write “Unclear”.
Use this format:
# Meeting Prep: [Meeting Title]
## Date
[Date or Unclear]
## Time
[Time or Unclear]
## Attendees
- [Attendee names or Unclear]
## Objective
- [Main objective]
## Context
- [Relevant background]
## Previous status
- [Known previous status or Unclear]
## Likely topics
- [Topic 1]
- [Topic 2]
- [Topic 3]
## Questions to ask
- [Question 1]
- [Question 2]
- [Question 3]
## Decisions needed
- [Decision 1]
- [Decision 2]
## Risks / sensitive points
- [Risk 1]
- [Risk 2]
## Practical angle to bring into the meeting
- [Short practical recommendation]
Save meeting prep files under:
~/Meetings/Prep/YYYY/MM/YYYY-MM-DD - [Meeting Title].md
Add these standing instructions to my OpenClaw workspace:
When I ask you to process a meeting transcript, convert it into structured meeting notes.
Rules:
1. Do not invent facts.
2. Do not invent decisions.
3. Do not invent deadlines.
4. If a deadline is missing, write “Unclear”.
5. Keep the notes concise, structured, and practical.
6. Extract action items in this exact format:
[Owner] - [Task] - [Due date]
7. Save the final notes automatically to the correct folder.
Use this format:
# Meeting Notes: [Meeting Title]
## Executive summary
- [Short summary]
## Key discussion points
- [Point 1]
- [Point 2]
- [Point 3]
## Decisions
- [Decision 1]
- [Decision 2]
## Action items
- [Owner] - [Task] - [Due date]
## Risks / open questions
- [Risk or open question]
## Next follow-up
- [Follow-up item]
Save post-meeting notes under:
~/Meetings/Notes/YYYY/MM/YYYY-MM-DD - [Meeting Title].md
I have a meeting to plan how learning OpenClaw can improve our team’s efficiency and output.
Prepare me for this meeting using my standard meeting prep format.
Process this meeting transcript using my standard meeting notes format.
Meeting title:
Learning OpenClaw to Improve Team Efficiency and Output
Transcript:
[paste transcript here]
After processing:
1. Extract the executive summary.
2. Extract key discussion points.
3. Extract decisions.
4. Extract action items in owner-task-due date format.
5. Mark missing due dates as “Unclear”.
6. Save the final notes to the correct meeting notes folder.
Create a file in my OpenClaw workspace called:
Linux_Research_Methodology.md
Add the following methodology:
# Linux Research Methodology
## Objective
Produce creator-friendly Linux research that helps me make accurate, useful, and interesting YouTube videos.
Focus on:
- what actually changed
- why it matters to desktop Linux users
- what is genuinely new
- what is overhyped
- what the community is saying
- what practical users should know
Avoid generic AI fluff.
## Source priority
### 1. Official sources first
Always check official sources first:
- distro release notes
- official blog posts
- changelogs
- GitHub/GitLab releases
- package documentation
- developer mailing lists if relevant
### 2. Trusted secondary sources
Then check trusted Linux/open-source news sources:
- Phoronix
- OMG! Ubuntu
- 9to5Linux
- The Register
- LWN if available
- official project forums
### 3. Community reaction
Then check community sentiment:
- Reddit
- Linux forums
- distro communities
- Hacker News if relevant
- GitHub issues if relevant
## Output format
When researching a Linux topic, use this structure:
# [Topic]
## 1. What changed
- List the real technical changes.
## 2. Why it matters
- Explain why desktop Linux users should care.
## 3. What people like
- Summarize positive community reaction.
## 4. What people dislike
- Summarize complaints, concerns, bugs, or pushback.
## 5. Does the hype match reality?
- Give a direct, balanced answer.
## 6. Creator-friendly summary
- Give me a short script-ready summary I can use in a video.
## Rules
- Be specific.
- Prefer concrete examples over vague claims.
- Separate fact from opinion.
- Do not blindly repeat marketing language.
- Mention uncertainty when sources are weak.
- Keep the final output useful for a Linux YouTube video.
Add this standing instruction to my main OpenClaw instructions:
For any Linux distro, Linux desktop, Linux software, open-source, or Linux news research, always read and follow Linux_Research_Methodology.md before searching the web or producing the final answer.
Ubuntu 26.04 LTS: does the hype match reality?
Follow my Linux research methodology.
Give me a creator-friendly report.
From now on, whenever I start a message with “Remember:” or “Remember that”, immediately save the key information permanently to the correct memory file.
Use:
- MEMORY.md for long-term facts, ideas, preferences, principles, and reusable knowledge.
- Daily log for short-term or date-specific notes.
Rules:
1. Save the information without asking follow-up questions unless the note is unclear.
2. Keep the saved memory concise.
3. Preserve the original meaning.
4. Reply only with a short confirmation like “Saved.”
5. When useful, add a lightweight category such as Idea, Workflow, Quote, Preference, or Research.
Remember: Tiago Forte says a second brain turns raw info into actionable output.
Remember: My best creative ideas hit during phone-free evening walks.
Remember: Research shows connected knowledge improves retention by around 65%.
Remember: Video idea - How OpenClaw makes your Obsidian vault actually think for you.
Remember: Experiment with daily voice notes via WhatsApp into OpenClaw.
Synthesize all five memories into a practical strategy for my YouTube content and personal workflow.
I want:
1. A core strategy.
2. Content ideas for my YouTube channel.
3. A personal workflow I can repeat weekly.
4. Any useful warnings or things I should verify.
5. A short video sequence I can build from these memories.
Set up my development environment on this Linux machine.
Goal:
Install and verify:
- Visual Studio Code
- Git
- Node.js
- npm
Important constraints:
1. First detect the Linux distro and package manager.
2. Use apt/deb packages on Ubuntu/Debian-based systems.
3. Do not install VS Code as a Snap.
4. Install VS Code using Microsoft’s official .deb package or official apt repository.
5. If Node.js is already installed, do not reinstall it. Just tell me the installed version.
6. If Git is already installed, do not reinstall it. Just tell me the installed version.
7. For any privileged command, use the system graphical authentication dialog with pkexec.
8. Never ask me to type my sudo password in chat.
9. Do not make unrelated system changes.
10. After installation, verify everything by printing:
- code --version
- git --version
- node --version
- npm --version
11. When everything is done, launch Visual Studio Code.
Before making changes:
- briefly tell me what you detected
- tell me which tools are already installed
- tell me what you are going to install
Then do the work.
Create a local Linux diagnostic bundle generator script.
Script path:
~/.openclaw/workspace/scripts/linux-diagnostic-bundle.sh
Make the script executable.
Goal:
The script should collect useful troubleshooting information from this Linux system and save it into a timestamped folder on my Desktop.
Output folder format:
~/Desktop/linux-diagnostic-bundle-YYYY-MM-DD_HH-MM-SS
The script must create these subdirectories:
- system
- storage
- network
- packages
- logs
Collect the following information:
system/
- OS information
- kernel version
- uptime
- CPU information
- memory information
- logged-in user/session information where safe
storage/
- lsblk output
- df -h output
- mounted filesystems
network/
- network interfaces
- IP addresses
- routing table
- DNS configuration if available
packages/
- apt package summary if apt exists
- flatpak list if flatpak exists
- snap list if snap exists
logs/
- failed systemd services
- recent journalctl errors
- recent dmesg warnings/errors
Also create:
summary.md
The summary should include:
- distro
- kernel
- CPU model
- total memory
- disk usage summary
- failed services count
- whether demo-safe mode was enabled
Safety rules:
1. The script must be read-only.
2. Do not modify system settings.
3. Do not install packages.
4. Do not delete files.
5. Do not require sudo.
6. Skip missing commands gracefully.
7. If a command is unavailable, write a short note into the relevant output file.
8. Use safe quoting.
9. Handle spaces in paths.
10. Print a clear final message showing where the bundle was created.
Add a -demo-safe flag.
When -demo-safe is used, redact sensitive information:
- usernames
- hostnames
- home directory paths
- MAC addresses
- public IP addresses if detected
After creating the script:
1. Make it executable.
2. Run it once in demo-safe mode.
3. Show me:
- the script path
- the exact command used to run it
- the generated output folder
- a tree view of the generated bundle
- a short summary of what was collected
Create a fresh Ubuntu setup script for my personal post-install routine.
The script should:
1. Update apt package lists.
2. Install my commonly used apps.
3. Configure basic GNOME settings.
4. Configure dock behavior.
5. Set my preferred DNS if possible.
6. Skip anything already configured.
7. Ask before making major changes.
8. Print a final summary.
Make it safe, readable, and easy to edit.
Do not run it automatically.
Save it to:
~/.openclaw/workspace/scripts/fresh-ubuntu-setup.sh
Create a shell script that generates my standard YouTube video project folder structure.
The script should:
1. Ask me for the video title.
2. Convert the title into a safe folder name.
3. Create the project folder.
4. Inside it, create:
- audio
- screen-recordings
- camera-recordings
- raw-footage
- notes
- thumbnails
- exports
- assets
5. Create a README.md with the project title and creation date.
6. Never overwrite an existing project folder.
7. If a folder already exists, safely create a numbered variant.
Save the script to:
~/.openclaw/workspace/scripts/create-video-project.sh
Make it executable but do not run it yet.
Create a shell script that watches a folder for new screen recordings and compresses them automatically.
Requirements:
1. Watch this folder:
~/Videos/Recordings-To-Compress
2. Output compressed videos to:
~/Videos/Compressed-Recordings
3. Use ffmpeg.
4. Preserve the original files.
5. Skip files that are still being written.
6. Compress to a good YouTube-friendly format.
7. Add logs to:
~/Videos/Compressed-Recordings/compression-log.md
8. Handle spaces in filenames.
9. Skip files that were already processed.
10. Print a clear summary.
Save the script to:
~/.openclaw/workspace/scripts/screen-recording-compressor.sh
Make it executable.
Do not run it automatically until I ask.