Various Tips and How To Dos…

Table of Contents:

How to Reduce PDF File Size
How to Resize Video Recording
How to Bootstrap a New GitHub Project
How to use GitHub Templates
How to use GitHub Actions
How to Update Claude Code

How to Reduce PDF File Size

To reduce the size of a PDF from the command line—especially one made of heavy scanned images—the industry standard tool you want is Ghostscript (gs - Ghostscript is a tool for creating, viewing, and transforming PostScript and PDF documents).

If you don’t have it installed, you can easily grab it via Homebrew:

brew install ghostscript

Once installed, here is the command to compress your scanned PDF:

gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dPDFSETTINGS=/ebook -dNOPAUSE -dQUIET -dBATCH -sOutputFile=compressed_output.pdf original_scanned.pdf

What makes this work:

The magic happens with the -dPDFSETTINGS flag, which dictates the compression level.

/ebook: (150 dpi) This is usually the sweet spot for scanned documents, offering a much smaller file size while keeping the text perfectly readable.
/screen: (72 dpi) Use this if you need the absolute smallest file possible and don’t mind a drop in visual quality.
/printer: (300 dpi) Use this if you need to retain high quality for physical printing, though it won’t compress as aggressively.

The /ebook tag itself is a fixed preset. It acts as a shortcut that tells Ghostscript to apply a whole bundle of internal settings, which happen to include downsampling images to exactly 150 DPI. You can’t pass a custom DPI directly into the /ebook tag (like /ebook:200 or something similar).

However, if you want complete control to dial in a very specific DPI—say you want 200 DPI because 150 is a little too blurry, or 100 DPI because you need it even smaller—you absolutely can.

To do this, you just drop the -dPDFSETTINGS preset and manually tell Ghostscript what resolution to use for the images inside the PDF.

Here is how you would run the command to force a custom DPI (using 200 as an example):

gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dNOPAUSE -dQUIET -dBATCH \
  -dDownsampleColorImages=true -dColorImageResolution=200 \
  -dDownsampleGrayImages=true -dGrayImageResolution=200 \
  -dDownsampleMonoImages=true -dMonoImageResolution=200 \
  -sOutputFile=custom_dpi_output.pdf original_scanned.pdf

What changed in this command:

Instead of relying on a one-size-fits-all preset, we are explicitly turning on downsampling for all image types (Color, Grayscale, and Monochrome) and manually setting the resolution for each to 200. You can change that 200 to whatever number fits your exact needs.

How to Resize Video Recording

FFmpeg is the industry standard and will do exactly what you want in seconds.

Install FFmpeg:

Mac: brew install ffmpeg
Windows: winget install ffmpeg

Command A: Letterbox (Fit inside box with black bars)

ffmpeg -i input.mp4 -vf "scale=1024:1024:force_original_aspect_ratio=decrease,pad=1024:1024:(ow-iw)/2:(oh-ih)/2" -c:a copy output_padded.mp4

Command B: Center Crop (Fill box, cut sides)

ffmpeg -i input.mp4 -vf "crop=1020:1020,scale=1024:1024" -c:a copy output_cropped.mp4

(Note: We crop to 1020 first because that is your video’s native height, then scale slightly to 1024).

Here is the exact command to scale the width to 1024 while keeping the aspect ratio automatically:

ffmpeg -i input.mp4 -vf scale=1024:-2 -c:a copy output_1024.mp4

Breakdown of the command:

-vf scale=1024:-2: This is the magic part.
- 1024: Sets your fixed width.
- -2: Tells FFmpeg to automatically calculate the height to maintain the aspect ratio AND ensure the height is an even number (divisible by 2). This is critical because many codecs (like H.264) will fail if you try to use an odd number for the height (e.g., 765 pixels).
-c:a copy: Copies the audio without re-encoding it. This makes the process much faster and prevents any audio quality loss.

Result: Your new video will likely be 1024x766.

How to Bootstrap a New GitHub Project

Your logic is spot on - it’s essentially “forking” the project manually to give yourself a clean slate while keeping the history if you want it (or clearing it if you don’t).

Here is the step-by-step workflow to get your new repository linked to your existing codebase.

1. Clone the Existing Project

First, clone the original repository into a new folder on your Mac.

git clone https://github.com/username/old-repo.git my-new-project
cd my-new-project

2. Create the New Repo on GitHub

Go to GitHub and create a New Repository.
Give it your new name (e.g., my-new-project).
Important: Do not initialize it with a README, license, or .gitignore, as you already have these in your local clone.
Copy the new repository’s URL (e.g., https://github.com/username/my-new-project.git).

3. Update the “Remote” Origin

Currently, your local folder thinks it should push to the old repository. You need to point it to the new one.

Check current remote:
```
git remote -v
```

Change the remote URL:

git remote set-url origin https://github.com/username/my-new-project.git

Verify the change:
```
git remote -v
```
(It should now show your new repo URL).

4. Optional: Clean Slate vs. Keeping History

Decide if you want to keep the commit history from the old project.

Option A: Keep History (Standard) Just push your existing branches to the new repo:
```
  git push -u origin main
```

Option B: Fresh Start (Clean History) If you want the new project to start at “Initial Commit” with no record of the old work:

  rm -rf .git
  git init
  git add .
  git commit -m "Initial commit for new project"
  git remote add origin https://github.com/username/my-new-project.git
  git push -u origin main

5. Final Housekeeping

Since you are starting a new project, don’t forget to update the following files if they exist:

package.json or requirements.txt: Update the project name and version.
README.md: Update the titles and descriptions.
Environment Variables: If you have a .env.example, make sure it reflects any new configurations.

How to use GitHub Templates

GitHub’s Template Repository feature is the most efficient way to manage “boilerplate” code. Instead of manually cloning and changing remotes every time, you can turn any existing repository into a blueprint for future ones.

1. How to Turn Your Project into a Template

Navigate to your existing repository on GitHub.com.
Click on the Settings tab at the top.
In the General section, look for the checkbox labeled Template repository.
Check that box. GitHub will automatically save the change.

2. How to Use the Template

Once a repo is marked as a template, a new green button appears at the top of the repository page: “Use this template”.

Clicking “Use this template” allows you to “Create a new repository.”
The Benefit: It creates a brand-new repo with a single “Initial commit” containing all your files, but without the entire commit history of the original project. This keeps your new project’s history clean from the start.

3. Advanced: Using “Repository Dispatch” for Automation

If your projects involve complex setups (like setting up an Oracle DB schema or configuring a specific Python environment), you can use GitHub Actions with your template.

You can include a .github/workflows/post-create.yml file in your template. When you create a new repo from that template, the action can automatically:

Rename variables in your README.md.
Update the package.json or pyproject.toml with the new repo name.
Run a “Sanity Check” test suite to ensure the boilerplate is working.

4. Comparison: Cloning vs. Templating

Feature	Manual Clone/Remote Change	GitHub Template
History	Keeps all old commits (unless deleted).	Starts with a clean “Initial commit.”
Speed	Requires Terminal commands.	One click on the GitHub UI.
Automation	Manual search/replace of names.	Can be automated via GitHub Actions.
Visibility	Works for any repo you can access.	Must be explicitly marked as “Template.”

5. Using Template Repository in Development

Making a repository a Template does not change how it functions for you as a developer. You can still push code, create branches, merge Pull Requests, and run CI/CD pipelines exactly as you did before.

Think of the “Template” status as an additional superpower rather than a change in state.

How it behaves in practice:

Regular Development: You continue to work in your local folder, git push to the origin, and manage your files.
The “Snapshot” Effect: When you (or someone else) clicks “Use this template”, GitHub essentially takes a “snapshot” of the code exactly as it exists on your default branch (usually main) at that specific moment.
Future Updates: If you add a new Python utility script or a new Oracle SQL initialization file to your template repo tomorrow, any new projects created from that template from that point forward will include those updates. Existing projects you already bootstrapped will not be affected.

A Tip for Your Two New Projects

Since you are bootstrapping two projects at once, here is how to handle the “Template” workflow effectively:

Finalize the Core: Make sure your “Template” repo has the common denominator code both projects need (e.g., your MinIO connection logic or basic logging patterns).
Push to Template: Commit those changes to GitHub.
Spawn Project A: Click “Use this template” $\rightarrow$ Name it Project-A.
Spawn Project B: Click “Use this template” $\rightarrow$ Name it Project-B.
Specialize: Now, go into the local folders for A and B and add the specific agentic patterns or threat-intel modules unique to each.

One Small Caveat

If you use the GitHub Action to rename placeholders, remember that the Action runs in the new repositories (Project-A and Project-B), not in the template itself. This keeps your template “clean” with the placeholders intact for the next time you need it.

Pro-Tip for Python Projects

Since you often work with many individual Python scripts grouped by patterns, you might want to include a setup.py or pyproject.toml in your template. This allows you to run pip install -e . immediately after creating a new project so all your internal modules are correctly pathed.

How to use GitHub Actions

Since your projects often involve complex structures—like specific Python patterns, Oracle DB schemas, or threat intelligence modules—automating the “handover” from a template to a new repository saves a lot of manual renaming.

We can use GitHub Actions to watch for the repository_created event. When you create a new repo from your template, this script will automatically find placeholders and replace them with your new project’s name.

1. The Automation Script (`.github/workflows/post-create.yml`)

Create this file in your template repository. It uses a simple sed command to swap out a placeholder (like PROJECT_NAME_PLACEHOLDER) with the actual name of your new repo.

name: Post-Template Initialization
on: [push] # Fires on the first push (which happens during template creation)

jobs:
  setup-new-repo:
    if: github.run_number == 1 # Only runs once, right after creation
    runs-on: ubuntu-latest
    steps:
      - name: Checkout Code
        uses: actions/checkout@v4

      - name: Rename Placeholders
        run: |
          # Replace placeholder in README and configuration files
          find . -type f -not -path '*/.*' -exec sed -i "s/PROJECT_NAME_PLACEHOLDER/$/g" {} +
          
      - name: Commit and Push Changes
        run: |
          git config --global user.name "github-actions[bot]"
          git config --global user.email "github-actions[bot]@users.noreply.github.com"
          git add .
          git commit -m "chore: initialize project from template"
          git push

2. How to Set Up Your Template Files

To make the script above work, you just need to use the string PROJECT_NAME_PLACEHOLDER anywhere you want the new name to appear.

README.md:

  # PROJECT_NAME_PLACEHOLDER
  This project was generated from a custom pattern template.

setup.py or pyproject.toml:

  name="PROJECT_NAME_PLACEHOLDER",
  version="0.1.0",

3. Workflow Comparison: Template vs. Manual

Step	Manual Method	Automated Template Method
Creation	`git clone` + `remote set-url`	Click “Use this template”
History	Full history remains	Clean “Initial commit”
Renaming	Manual search and replace	Handled by GitHub Action
Dependencies	Manual `pip install`	Action can pre-verify environment

Why this fits your current projects

Since you are testing different agentic design patterns and building a threat intelligence network, you likely have a specific directory structure you prefer (e.g., /scripts, /data, /logs). By using this template:

Your MinIO or ChromaDB connection strings can stay consistent.
Your Oracle DB initialization scripts stay in a known location.
Every new experiment starts with the same clean, professional structure without you typing a single mkdir command.

How to Update Claude Code

# Check for any Claude Code (Node.js) processes
ps aux | grep -i claude | grep -v grep

# Also check for any node processes that might be Claude Code related
ps aux | grep -i "claude-code" | grep -v grep

# Check for the caffeinate process that Claude Code spawns (prevents Mac from sleeping)
ps aux | grep caffeinate | grep -v grep

If you see any results, you can kill them with kill <PID> or, to be thorough:

# Kill all Claude Code related processes
pkill -f "claude-code"
pkill -f caffeinate  # only if it was spawned by Claude Code

2. Upgrading Claude Code

Here’s the full picture:

1. Confirm no Claude processes are running:

# Check for any Claude Code processes
ps aux | grep -i claude | grep -v grep

# Check for caffeinate (Claude Code spawns this to prevent Mac from sleeping)
ps aux | grep caffeinate | grep -v grep

If anything shows up, kill it with kill <PID> or pkill -f "claude-code".

2. Upgrade Claude Code:

The recommended approach depends on how you originally installed it. The npm installation method is now deprecated — the native installer is faster, requires no dependencies, and auto-updates in the background.

If you installed via npm (most likely given your setup), migrate to the native installer:

# Install the native binary
curl -fsSL https://claude.ai/install.sh | bash

# Remove the old npm installation
npm uninstall -g @anthropic-ai/claude-code

You can also run claude install from an existing npm installation to install the native binary alongside it, then remove the npm version.

If you’re already on the native installer, it auto-updates in the background. To force an immediate update:

claude update

After upgrading, verify the installation:

claude --version
claude doctor

claude doctor will verify the installation type and version.

One note on release channels: You can control which release channel Claude Code follows for auto-updates with the autoUpdatesChannel setting: “latest” (the default, new features immediately) or “stable” (about one week behind, skipping releases with major regressions). You can configure this via /config → Auto-update channel inside Claude Code.

Igor's KB Site

A personal website for infomation sharing

Various Tips and How To Dos…

How to Reduce PDF File Size

What makes this work:

What changed in this command:

How to Resize Video Recording

Breakdown of the command:

How to Bootstrap a New GitHub Project

1. Clone the Existing Project

2. Create the New Repo on GitHub

3. Update the “Remote” Origin

4. Optional: Clean Slate vs. Keeping History

5. Final Housekeeping

How to use GitHub Templates

1. How to Turn Your Project into a Template

2. How to Use the Template

3. Advanced: Using “Repository Dispatch” for Automation

4. Comparison: Cloning vs. Templating

5. Using Template Repository in Development

How it behaves in practice:

A Tip for Your Two New Projects

One Small Caveat

Pro-Tip for Python Projects

How to use GitHub Actions

1. The Automation Script (`.github/workflows/post-create.yml`)

2. How to Set Up Your Template Files

3. Workflow Comparison: Template vs. Manual

Why this fits your current projects

How to Update Claude Code

2. Upgrading Claude Code

Various Tips and How To Dos…

How to Reduce PDF File Size

What makes this work:

What changed in this command:

How to Resize Video Recording

Breakdown of the command:

How to Bootstrap a New GitHub Project

1. Clone the Existing Project

2. Create the New Repo on GitHub

3. Update the “Remote” Origin

4. Optional: Clean Slate vs. Keeping History

5. Final Housekeeping

How to use GitHub Templates

1. How to Turn Your Project into a Template

2. How to Use the Template

3. Advanced: Using “Repository Dispatch” for Automation

4. Comparison: Cloning vs. Templating

5. Using Template Repository in Development

How it behaves in practice:

A Tip for Your Two New Projects

One Small Caveat

Pro-Tip for Python Projects

How to use GitHub Actions

1. The Automation Script (.github/workflows/post-create.yml)

2. How to Set Up Your Template Files

3. Workflow Comparison: Template vs. Manual

Why this fits your current projects

How to Update Claude Code

1. Confirming no Claude-related processes are running on your Mac

2. Upgrading Claude Code

1. The Automation Script (`.github/workflows/post-create.yml`)