experiment-design-kit

Looking at your skill's focus on structured experimentation—I'm curious how you're handling the balance between prescriptive guidance and letting users adapt the framework to their specific research contexts.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 70/100, C territory. This is based on Anthropic's Progressive Disclosure Architecture standards for agentic skills. Your strongest area is Writing Style (8/10)—clear, instructional voice without marketing fluff. The weakest? Spec Compliance (11/15) and Progressive Disclosure (21/30), mainly because your description lacks trigger phrases and you're embedding templates instead of referencing them.

What's Working Well

• Solid operational structure – Your 5-step Framework (hypothesis → variants → guardrails → measurement → learning loop) is genuinely useful and gives people a concrete path forward
• Good terminology consistency – You stick with "hypothesis," "guardrail," and "variant" throughout without switching synonyms around, which keeps things scannable
• Practical guardrail examples – The KPI Guardrails section (activation ±3%, revenue −2%, support tickets <5%) grounds the framework in real numbers rather than abstract concepts
• Clear cross-functional thinking – Your Operating Model acknowledges different roles (Product, Data, GTM) which reflects real team structures

The Big One: Missing Trigger Phrases in Description

This is your highest-leverage fix. Right now your description says:

"Toolkit for structuring hypotheses, variants, guardrails, and measurement plans."

That's generic enough that Claude won't activate this skill when someone asks "help me design an A/B test" or "create experiment brief."

**Here's the...