Feedback on your deps skill

[RichardHightower]

I took a look at your deps skill and wanted to share some thoughts.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 77/100, which lands you in C territory – solid fundamentals but some gaps to close. Based on Anthropic's best practices, your Utility pillar is strong (19/20) with real problem-solving power for dependency upgrades. But Spec Compliance is a hard zero (0/15), which is why the grade dips. The good news? This is fixable.

What's Working Well

• Excellent practical value – You're solving a real problem (dependency upgrade analysis) with multi-platform support (Node, iOS, Android). That's not trivial.
• Solid trigger patterns – Your argument-hint shows concrete examples like "react 17 → 18" and "iOS 16" that make it clear what this skill does.
• Strong verification protocol – The V1-V4 verification steps with confidence levels (high/medium/low) and re-execution logic show you've thought through feedback loops.
• Clean architecture – The Progressive Disclosure structure (SKILL.md linking to 4 reference files) is well-organized and token-efficient at 28/30.

The Big One: Fix Your Frontmatter

Here's the deal – your SKILL.md frontmatter is invalid YAML, and that's tanking your entire Spec Compliance score. It's cut off mid-description:

---
name: deps
description: Analyze dependency usage for library/framework/SDK upgrades
model: sonnet
allowed-tools: Bash, Glob, Grep, Read, Write, WebSearch, WebFetch
argument-hint: [target, e.g., "r

That argument-hint line never closes. Fix it like this:

---
name: deps
description: Analyze dependency upgrade paths for libraries, frameworks, and SDKs. Use when asked to "analyze deps", "check upgrade path", "deps for react", or "iOS dependency analysis".
model: sonnet
allowed-tools: Bash, Glob, Grep, Read, Write, WebSearch, WebFetch
---

This single fix gets you +5 points immediately, plus your description now includes trigger phrases so people actually know when to invoke this. That's easily +9 points total on Spec Compliance.

Other Things Worth Fixing

1. Add trigger phrases to your description – Right now you're missing the "Use when asked to..." pattern. Add triggers like "analyze deps", "check upgrade path", "dependency upgrade" so users know when to call this.

2. Trim the workflow redundancy – Your Phase descriptions appear in both SKILL.md (lines 49-115) and workflow.md. In SKILL.md, reduce phases to one-liners linking to workflow.md: "Phase 1: Detect Current Version → See workflow.md". Saves tokens and improves navigation.

3. Add TOC to longer files – reference.md (472 lines) and verification-guide.md (297 lines) need tables of contents. Users shouldn't have to scroll blind through that much content.

4. Pair examples with inputs – Your output-template.md shows YAML output, but not the exact command that generated it. Add "Input: /deps 'react 17 → 18'" before each example so it's reproducible.

Quick Wins

• Fix frontmatter YAML syntax (+5 points) ⚡
• Add trigger phrases to description (+4 points)
• Remove workflow duplication (+1 point)
• Add TOCs to reference files (+1 point)

That's +11 points with focused edits – gets you to 88/100 (B territory) without major rework.

---

Checkout your skill here: [SkillzWave.ai](https://skillzwave.ai) | [SpillWave](https://spillwave.com) 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.