cold-email-personalization

I've been digging into your cold-email-personalization skill, and the way you're handling dynamic content generation is surprisingly thoughtful—most tools in this space either oversimplify it or make it too rigid. The 89/100 score reflects solid fundamentals with room to tighten up a couple of architectural choices, but this is genuinely useful work for the personalization problem space.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 89/100, solid B-grade territory. This is based on Anthropic's skill best practices. Your strongest areas are Utility (19/20) and Progressive Disclosure Architecture (28/30), which tells me the skill actually solves problems and is well-structured for token efficiency. Weakest spot is Spec Compliance (12/15)—nothing catastrophic, just some metadata gaps.

What's Working Well

• Layered structure is chef's kiss. Your SKILL.md clocks in at just 44 lines with 10 well-organized reference files. That's exactly how token economy should work—tight overview, depth on demand. The way you buried the 3:1 recipient-to-sender ratio and word count targets in dedicated files means users only load them when they need that specificity.

• Real validation loops. You've got the "Would I Reply?" test, a 0-100 scoring rubric, and a QA checklist that actually prevents hallucination. Most skills just hand you a template and wish you luck; you're teaching people how to verify.

• Five campaign examples with context. Not just templates—you're showing why each one works (tech startup cold email hits different than enterprise). That's the difference between a reference and something people actually use.

The Big One: Missing Trigger Terms

Your description is **too vague...