---
title: "excel-ops"
description: "Read, clean, analyse and generate Excel workbooks without the silent wrong answers, header detection, numbers stored as text, formula cells, date timezone drift, phantom rows, and styled report generation. Ships a determ"
canonical: https://agentpluginsdirectory.com/plugins/excel-ops
last-updated: 2026-09-21
---

# excel-ops
Read, clean, analyse and generate Excel workbooks without the silent wrong answers, header detection, numbers stored as text, formula cells, date timezone drift, phantom rows, and styled report generation. Ships a deterministic benchmark.
- Slug: excel-ops
- Publisher: AgentUI
- Repository: https://github.com/agentui-ai/excel-ops
- Manifest: plugin.json
- Version: 0.1.0
- License: MIT
- Category (editorial): other
- Skills: 1 (excel-ops)
- MCP servers: 0
- Stars: 0
- Repository created: 2026-09-19
- Repository last pushed: 2026-09-19
- Publisher type: Organization
- Listing: https://agentpluginsdirectory.com/plugins/excel-ops
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## What excel-ops does, in the publisher's words

excel-ops

An agent skill for Excel work that does not quietly return the wrong number, plus a benchmark that proves it.

Spreadsheet code rarely crashes. It returns a plausible number that is wrong, and nobody notices until the month closes. excel-ops teaches a coding agent the nine failure modes that cause it, ships reference implementations for reading and generating workbooks, and ships a benchmark that runs the from-memory version side by side with the skill's against real.xlsx files.

From the project README, punctuation lightly normalized. Full text: https://raw.githubusercontent.com/agentui-ai/excel-ops/HEAD/README.md

## Skills

- excel-ops: Read, clean, analyse and generate Excel workbooks (.xlsx/.csv) without the silent wrong answers. Use when the user has a spreadsheet to parse, summarise, validate, join, convert or produce; when a total from a sheet looks off; when dates shift by a day, numbers arrive as text, formulas read back as…

Descriptions come from the frontmatter of each SKILL.md, punctuation lightly normalized.