request-analyzer

I took a quick look at how you're parsing and validating HTTP requests—the way you're handling edge cases in the request body analysis is solid. With a 92/100, you're in that interesting space where the fundamentals are clearly dialed in; I'd be curious what pushed a couple points off.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 92/100, solid A territory. This is based on Anthropic's best practices for skill design. Your strongest area is Writing Style (9/10)—the voice is consistent and objective throughout. Weakest link is Spec Compliance (12/15), mostly because you're only using 1-2 trigger phrases when you could squeeze out more discoverability.

What's Working Well

• Layered architecture is chef's kiss. SKILL.md stays clean as an overview, with detailed analysis logic split across buzzword-dictionary.md and classification-criteria.md. That's exactly the token-efficient separation that matters.
• Your five-step framework is rock solid. The progression from intent detection → complexity assessment → validation is intuitive and practical. The tables make it scannable.
• Trigger terms are well-chosen. "Classify requests", "prioritize work", "route to specialists"—these actually describe what developers need, not buzzword soup.

The Big One: Missing Table of Contents

All three files exceed 100 lines but don't have TOCs. For SKILL.md (149 lines) and your reference files (113 and 144 lines), this hurts navigation and costs you 2-3 points in Progressive Disclosure.

Fix: Add a simple TOC at the top of each file:

# Request Analyzer

## Table of Contents
- [When to Use](#when-to-use)
- [Analysis Framework](#analysis-framework)
- [Output Format](#output-format)
- [Examples](#examples)

Classifies incoming requests...

This alone bumps you +2 points toward 94/100.

Other Things Worth Fixing

1. Examples are too chatty (SKILL.md lines 104-149). You're explaining the analysis to Claude, but Claude doesn't need the narrative. Trim those down to just the input→output format you define. Saves tokens, gains +1 point.

2. Step 5 inconsistency — Earlier steps use tables, but Step 5 has checklists. Pick one format and stick it. Minor, but +2 points for workflow clarity.

3. Bump description trigger phrases. You've got solid ones, but add 1-2 more (like "request prioritization" or "dependency validation"). Helps with discoverability.

Quick Wins

• Add TOCs to all three files → +2 points
• Tighten example narrative → +1 point
• Standardize Step 5 format → +2 points
• Add 1-2 trigger phrases → marginal, but improves SEO

You're legitimately close to 95+. These are all small structural tweaks, not content problems.

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.