vague

I've been digging through skills lately and yours came up with a solid 90—definitely above the noise floor. The concept feels intentionally vague in a way that suggests you're handling some real complexity under the hood, which is exactly when clarity becomes the hard part.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

TL;DR

You're at 90/100, solid A territory. This is based on Anthropic's skill best practices rubric. Strongest area: Utility (19/20)—you've actually solved a real problem with comprehensive test data generation. Weakest area: Spec Compliance (12/15)—your description could use a few more trigger phrases to help people discover this when they need it.

What's Working Well

• Progressive Disclosure is chef's kiss. Your SKILL.md stays lean and focused while references like syntax.md and plugins.md handle the deep dives. That's exactly how layering should work (28/30 on PDA).
• The utility is genuinely strong. You're not just documenting a tool—you're providing real validation patterns (constraints, dataset validation, OpenAPI validation). That's the difference between a reference and something people actually use.
• Token economy. Nearly every sentence pulls its weight. No fluff about "why Vague matters" or "test data is important"—just straight to solving the problem.
• Consistent terminology throughout. Schema, dataset, superposition, constraints—you stick with it, which keeps things clear even when diving into complex features.

The Big One: Missing End-to-End Workflow

Here's what's holding you back from a 91+: Your Quick Start shows syntax, but there's no complete workflow showing the actual problem-solving loop. Right now someone can learn how to write Vague, but not when or why in context.

Fix this by adding to SKILL.md:

## Complet...