document-processing-xlsx

I took a look at your document-processing-xlsx skill and wanted to share some thoughts.

Links:

• GitHub
• Your skill in the SkillzWave marketplace

TL;DR

You're at 51/100 – that's F territory, which means this needs some real work before it's ready. This is based on Anthropic's skill best practices around architecture, usability, and utility.

Your strongest area is Spec Compliance (12/15) – the frontmatter is clean and the name format is solid. The biggest gaps are Progressive Disclosure Architecture (10/30) and Ease of Use (14/25). You're losing a lot of points on structure and workflow clarity.

What's Working Well

• Clean metadata – Your YAML frontmatter is valid with required fields, and the hyphen-case naming is correct
• Good trigger discovery – You've packed the description with relevant terms (Excel, xlsx, import, export, reports) that make the skill findable
• Realistic scope – The skill tackles a genuine problem that developers actually need solved

The Big One: Duplicate Sections & No Architecture

Here's what's holding you back the most: you've got two identical "When to use this skill" sections taking up lines 8-29, and the entire skill lives in one flat 42-line file with zero reference documents.

Why it matters: This wastes tokens and signals that the skill isn't organized for progressive disclosure. More critically, it tells users "here's everything you need" upfront, which violates how agentic skills should layer information.

The fix:

1. Merge the duplicate sections into one consolidated "When to use this skill" section
2. Create a references/ directory with:
   • references/library-comparison.md – Compare xlsx vs exceljs vs XLSX.js with decision criteria
   • references/common-patterns.md – Real workflows (read → transform → write with validation)
   • references/error-handling.md – File validation, corrupted data recovery, type checking

This structure shift alone could earn you +6-8 points.

Other Things Worth Fixing

1. Your example is incomplete – It shows creating a worksheet but doesn't show a full workflow. Add numbered steps: 1) Install library 2) Read file 3) Parse data 4) Validate & handle errors 5) Write output. This covers you on the "Ease of Use" front (+4 points).

2. No error handling or validation guidance – You mention xlsx/exceljs/SheetJS but don't address how to handle corrupted files, validate sheet names, or check data types. Add a "Common Pitfalls" section in the main file referencing your error-handling reference document (+4 points).

3. Single library bias – You describe multiple libraries but only demonstrate SheetJS. Either focus the description on one library with a note about alternatives, or add comparison criteria so users can pick the right tool (+2 points).

4. Terminology inconsistency – You flip between "Excel," "xlsx," and "spreadsheet" throughout. Pick one primary term and use it consistently.

Quick Wins (Ranked by Impact)

• Create reference files (+6 points) – This is the quickest path to better PDA scores
• Complete the workflow example (+4 points) – Add the full read→process→write cycle with error handling
• Remove duplicate section (+2 points) – Merge lines 8-29 into one cohesive section
• Add validation patterns (+4 points) – Show how to validate file format and data integrity
• Standardize terminology (+1 point) – Pick xlsx or Excel and stick with it

If you fix the architecture and workflow issues, you could realistically hit 65-70/100. The spec compliance piece is already solid, so you've got a good foundation to build from.

Checkout your skill here: SkillzWave.ai | SpillWave We have an agentic skill installer that install skills in 14+ coding agent platforms. Check out this guide on how to improve your agentic skills.