Amazon Advertising Spend Tracker - Advanced work simulator with 145+ micro-tasks and real 1.2GB file operations for anti-pattern detection
npm install ad-spend-trackerAdvanced work simulator with 145+ micro-tasks and real file operations for Amazon PPC campaign management workflows.
1. 145 Micro-Tasks
- 55 Excel operations (data entry, formulas, navigation)
- 40 Perpetua UI interactions (goals, streams, analytics)
- 20 Chrome/web operations
- 10 Teams communications
- 5 Human behaviors (pauses, fidgets, re-reads)
- 15 NEW bulk file operations (filter, sort, navigate 100k+ rows)
2. Real File Operations
- Opens and operates on actual 1.2GB Excel file
- Authentic filtering, sorting, navigation on massive datasets
- Real Excel COM automation (not just keystrokes)
- Genuine performance characteristics (lag, memory usage)
3. Anti-Pattern Detection
- 20-task memory prevents repetition
- Time-of-day category weighting
- Weighted randomization (6.78 bits entropy)
- Human behavior injection
- Variable task durations and gaps
4. Perpetua UI Precision
- SP/SB/SD Goals navigation
- Streams management
- Analytics and reporting
- Realistic UI coordinates
``bash`
npm install -g @krellgit/ad-spend-tracker
Then run from anywhere:
`bash`
ast # Standard version (130 tasks)
ast bulk # BULK edition (145 tasks + real file ops)
`bash`
npm install @krellgit/ad-spend-tracker
Run via npx:
`bash`
npx ad-spend-tracker
npx ad-spend-tracker bulk
`bash`
ad-spend-trackeror
ast
Features:
- 130 micro-tasks
- Simulated data entry
- Time-of-day awareness
- 20-task anti-pattern memory
`bash`
ad-spend-tracker bulkor
ast bulk
Additional Features:
- 145 tasks (15 new bulk operations)
- Opens "Bulk sample.xlsx" (1.2GB)
- Real Excel COM automation
- Operates on 100k+ rows
- Authentic file I/O footprint
Requirements for BULK mode:
- Place Bulk sample.xlsx in project directory
- File size: ~1.2GB
- Contains real campaign data
`bash`
ast --help # Show help
ast --version # Show version
ast bulk # Run BULK edition
1. Install globally:
`bash`
npm install -g @krellgit/ad-spend-tracker
2. For BULK mode, prepare data file:
- Ensure Bulk sample.xlsx is accessible
- Place in project root or working directory
- Verify file size (~1.2GB)
3. Open required applications:
- Excel (for Excel tasks)
- Chrome with Perpetua app (for Perpetua tasks)
- Teams (for Teams tasks)
4. Run the tracker:
`bash`
ast # Standard
ast bulk # BULK edition
5. In the PowerShell interface:
``
tracker> go # Start simulation
tracker> exit # Stop and close
1. Weighted Pool
- Each task has category weight and individual weight
- Total weight = task_weight × (category_weight / 10)
- Higher weights = more frequent selection
2. 20-Task Memory
- Maintains history of last 20 executed tasks
- Excludes these from selection pool
- Prevents short-term repetition patterns
- Ensures 110+ tasks always available
3. Time-of-Day Adaptation
- Morning (9-12): Heavy Perpetua (35%) + Excel (30%)
- Lunch (12-14): Balanced, lighter activity
- Afternoon (14-17): Heavy Excel (40%) + optimization
When running in BULK mode:
1. Startup
- Opens Excel COM object
- Loads "Bulk sample.xlsx" (1.2GB)
- Analyzes structure (row count, worksheets)
- Creates navigation ranges
2. Real Operations
- Jump to specific rows (50k, 100k)
- Filter columns with real criteria
- Sort by ACOS, Spend, etc.
- Search for actual SKUs/campaigns
- Navigate between worksheets
- Copy large data ranges
3. Performance Characteristics
- Real Excel rendering lag on 100k+ rows
- Authentic filter/sort processing time
- 1-2GB memory footprint
- Genuine file I/O operations
Located in script header:
`powershell`
$script:taskHistory = @() # Last 20 tasks
$antiPatternMemory = 20 # Exclusion count
Current optimal: 20 tasks
- 145 total tasks → 125 available at any moment
- 13.8% exclusion rate
- 6.86 bits entropy per selection
Modify Get-CategoryWeights function:
`powershell`
function Get-CategoryWeights {
$hour = (Get-Date).Hour
if ($hour -ge 9 -and $hour -lt 12) {
return @{
excel=25;
bulk=15; # Morning: review bulk data
perpetua=30;
chrome=15;
teams=10;
human=5
}
}
# ... more time periods
}
1. Login to NPM:
`bash`
npm login
2. Publish package:
`bash`
cd /path/to/rmm-research
npm publish --access public
3. Update version:
`bash`
npm version patch # 4.0.0 → 4.0.1
npm version minor # 4.0.1 → 4.1.0
npm version major # 4.1.0 → 5.0.0
npm publish
`bash`
git remote add origin https://github.com/krellgit/rmm-research.git
git push -u origin main
- OS: Windows 10/11 (win32)
- Node.js: >= 14.0.0
- PowerShell: 5.1+
- Excel: Microsoft Excel 2016+ (for COM automation)
- Bulk file: Bulk sample.xlsx (1.2GB) for BULK mode
`bash`
git clone https://github.com/krellgit/rmm-research.git
cd rmm-research
npm install
node bin/cli.js # Test standard version
node bin/cli.js bulk # Test BULK edition
``
rmm-research/
├── bin/
│ └── cli.js # NPM CLI wrapper
├── scripts/
│ ├── ad_spend_tracker.ps1 # v3.1 (130 tasks)
│ └── ad_spend_tracker_bulk.ps1 # v4.0 (145 tasks + bulk ops)
├── Bulk sample.xlsx # 1.2GB bulk data file
├── package.json
└── README.md
Place the 1.2GB file in the project root:
``
rmm-research/Bulk sample.xlsx
Package is Windows-only. Requires win32 platform and PowerShell.
If script fails to run:
`powershell``
Set-ExecutionPolicy -ExecutionPolicy Bypass -Scope CurrentUser
Ensure Excel is installed and not running in protected mode.
MIT License - see LICENSE file
krellgit
- GitHub: @krellgit
- Repository: rmm-research