excel-controller — OpenClaw Skill

RPA-style Excel automation for AI agents. Give your OpenClaw agent full control over the live Excel application — open workbooks, write data, create dashboards, run macros, and do everything a human can do, but faster.

---

What This Skill Does

Unlike file-based libraries (openpyxl, pandas) that manipulate .xlsx files on disk, excel-controller drives the running Excel application in real time — the same way a human would click and type:

• Open, create, save, and close workbooks
• Navigate sheets, select cells, read and write data
• Write formulas that Excel calculates natively
• Format cells: fonts, colors, borders, number formats, alignment
• Create charts: bar, line, pie, combo, radar, scatter
• Build pivot tables with custom fields and aggregations
• Add conditional formatting, data validation, and dropdowns
• Run existing VBA macros or inject and execute new ones at runtime
• Build complete multi-sheet dashboards with KPI cards
• Export sheets to PDF
• Sort, filter, and auto-fit columns

Platform support:

Platform	Engine	Package
Windows	COM Automation (win32com.client)	pywin32
macOS	xlwings bridge	xlwings
Linux	Not supported	Use excel-master instead

---

Prerequisites

• Microsoft Excel (or Microsoft 365) must be installed
• Python 3.8+

---

Installation

1. Clone or download

git clone https://github.com/3bslam/excel-controller-openclaw-skill.git

2. Copy the skill into OpenClaw

# Per-workspace (recommended)
cp -r excel-controller-openclaw-skill/excel-controller \
      ~/.openclaw/workspace/skills/excel-controller

# Or shared across all agents
cp -r excel-controller-openclaw-skill/excel-controller \
      ~/.openclaw/skills/excel-controller

3. Install Python dependencies

# Windows (full COM automation — recommended)
pip install pywin32

# macOS (xlwings bridge)
pip install xlwings

# Optional: data manipulation helpers
pip install pandas openpyxl

4. (Windows only) Enable VBA trust for macro injection

If you want the agent to inject and run VBA code at runtime:

Excel → File → Options → Trust Center → Trust Center Settings → Macro Settings → ☑ Trust access to the VBA project object model

5. Reload OpenClaw

/new
# or
openclaw gateway restart

---

Usage

Once installed, just describe what you want in natural language:

"Open Excel and create a sales dashboard with bar charts and KPI cards"
"Open my quarterly report and add a pivot table by region"
"Format the header row — dark blue background, white bold text"
"Create a line chart showing the revenue trend across Q1–Q4"
"Run the CalculateAll macro in this workbook"
"Add conditional formatting: green above 1000, red below 500"
"Export the Dashboard sheet to PDF"

The agent uses scripts/excel_safe.py automatically to manage the COM session safely.

---

Skill Configuration (optional)

In ~/.openclaw/openclaw.json:

{
  "skills": {
    "entries": {
      "excel_controller": {
        "enabled": true
      }
    }
  }
}

---

Repository Structure

excel-controller-openclaw-skill/
├── README.md                       # This file
├── LICENSE                         # MIT
├── .gitignore
└── excel-controller/
    ├── SKILL.md                    # Skill definition (frontmatter + full reference)
    ├── README.md                   # Quick-start guide
    └── scripts/
        └── excel_safe.py           # COM session manager + formatting helpers

scripts/excel_safe.py

A production-ready helper that provides:

Export	Description
excel_session(visible, display_alerts)	Context manager — handles COM init, Excel launch, and guaranteed cleanup
rgb(r, g, b)	Converts RGB to Excel's BGR integer format
COLORS	Professional color palette (header blue, input blue, positive green, etc.)
format_header(ws, range, ...)	One-call header formatting
format_currency(ws, range, fmt)	Apply currency number format
format_percentage(ws, range, decimals)	Apply percentage format
add_borders(ws, range, weight, color)	Add borders to a range
auto_fit(ws)	Auto-fit all columns and rows
write_data_block(ws, start_cell, data)	Bulk-write a 2D list (much faster than cell-by-cell)
disable_updates(excel)	Turn off screen update + manual calc for performance
enable_updates(excel)	Re-enable screen update + auto calc

Example:

from excel_safe import excel_session, write_data_block, format_header, rgb

with excel_session(visible=True) as excel:
    wb = excel.Workbooks.Add()
    ws = wb.ActiveSheet
    ws.Name = "Sales"

    headers = [["Product", "Q1", "Q2", "Q3", "Q4"]]
    data = [
        ["Product A", 15000, 18000, 17500, 21000],
        ["Product B", 12000, 13500, 14200, 16800],
    ]

    write_data_block(ws, "A1", headers + data)
    format_header(ws, "A1:E1")

    wb.SaveAs(r"C:\Users\me\Documents\sales.xlsx")

---

excel-controller vs excel-master

Feature	excel-controller	excel-master
Controls live Excel app	✅	❌
Requires Excel installed	✅	❌ (uses LibreOffice)
Real-time formula recalc	✅ Native	⚠️ LibreOffice headless
VBA macros	✅ Full	❌
Pivot tables	✅ Native	❌
Charts	✅ Full Excel	⚠️ openpyxl (limited)
Works on Linux	❌	✅
Headless / server use	⚠️ Background mode	✅
Speed on large files	⚠️ App overhead	✅ Fast

Use excel-controller when you need the full power of Excel — dashboards, pivot tables, VBA, real-time formulas, and interactive work on Windows/macOS.

Use excel-master when you need to create or process files on a server, in CI/CD, or on Linux.

---

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/my-improvement)
3. Commit your changes
4. Open a Pull Request

For bug reports and feature requests, please open an issue.

---

Community

This skill is a community contribution to the OpenClaw ecosystem. If you build something useful with it, share it — open an issue or discussion to show what you've made.

---

License

MIT — free to use, modify, and distribute.