drizzle

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

Links:

• GitHub
• Your skill on SkillzWave

The TL;DR

You're at 81/100, solid B-grade territory. This is based on Anthropic's skill best practices. Your strongest area is Ease of Use (22/25)—the metadata triggers and discoverability are really well done. Weakest area is Progressive Disclosure Architecture (24/30)—mainly because the file could be more efficiently layered and you're missing some navigation helpers for a ~270-line file.

What's Working Well

• Trigger phrases are sharp. You've got specific intent triggers like "create table" alongside artifact triggers like "pgTable" and even regex patterns for code detection. That's the kind of discoverability that actually helps agents find your skill.
• Templates are practical. The two template files (table-schema.ts.md and migration.ts.md) give concrete starting points, and your quick_start section includes real code examples with placeholders. People can actually copy-paste and iterate.
• Problem-solving is focused. You're tackling real Drizzle pain points: type exports, timezone handling, migration conflicts. Not abstract fluff—actual developer headaches.

The Big One: Missing Workflow Steps

Right now your quick_start shows code snippets but doesn't walk through the actual sequence. You jump from "here's Drizzle" to "here's a table definition" without the middle steps.

Why it matters: Developers using your skill want to know: fetch docs → create schema file → define table → export types → generate migration. That's the workflow.

The fix: Add a numbered workflow section before the code examples. Something like:

1. Use MCPSearch to fetch Drizzle docs
2. Create a new schema file (e.g., schema.ts)
3. Define your table using ...