turbo

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

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 79/100, which puts you in solid C territory — good fundamentals, but some gaps that are keeping you from reaching the next level. This is based on Anthropic's best practices for skill design. Your strongest area is Ease of Use (22/25) with excellent trigger coverage and metadata — the weakness is Progressive Disclosure Architecture (22/30), where you're losing points on scalability and token economy.

What's Working Well

• Trigger coverage is chef's kiss — You've got both intent phrases ("optimize monorepo builds") and artifact patterns (regex for turbo.json), plus CLI flag triggers. That's solid discoverability.
• Good practical examples — The JSON config examples showing common task patterns actually help Claude understand what you're solving for. Real workflows, not abstract theory.
• Clear constraints — Your "banned practices" and "required practices" sections give Claude guardrails. That's the kind of structure that prevents hallucinations.

The Big One: Step-by-Step Setup is Missing

Your quick start shows the turbo.json config but doesn't walk through the actual workflow. Right now Claude gets the destination but not the journey. This affects Ease of Use directly.

The fix: Add a numbered workflow before the config examples:

1) Verify turbo is installed (task turbo --version)
2) Create turbo.json in monorepo root
3) Define task pipeline (dependencies, outputs, cache rules)
4) Run turbo build --dry-run to validate
5) Execute full build with turbo build

This is +2-3 points right there.

Other Things Worth Fixing

1. That "Always fetch" phrasing in mcp_first — You're addre...