Create minimal probe plans to validate blocking product hypotheses.
npx skills add https://github.com/hujianbest/garage-agent --skill hf-experiment
Please help me install this Agent Skill. Skill: hf-experiment Source: https://github.com/hujianbest/garage-agent/tree/main/packs/coding/skills/hf-experiment Command: npx skills add https://github.com/hujianbest/garage-agent --skill hf-experiment
在进入 hf-specify 前（或处理 Blocking 假设尚未关闭时），为一条或少数几条关键假设产出一份结构化 probe plan：明确假设是什么、如何最小验证、多长时间、什么结果算通过、证据落在哪、结果如何回流。 本 Skill 的核心价值在于：把 discovery / spec 阶段模糊的"我们假设"变成可证伪的"一次小验证"，避免把未经检验的假设硬推给下游设计或实现。
Draft a minimal probe plan for your blocking hypothesis and run a quick validation before moving to hf-specify.
High-intent search queries and answers about installing and using this skill.
A probe plan validates a product hypothesis by defining its type, method, sample size, setup, and success/failure thresholds to generate falsifiable evidence before pushing mainline development.
A minimal probe plan in product discovery is a structured validation framework that transforms vague assumptions into falsifiable tests with defined timeboxes, evidence paths, and success thresholds to prevent unverified hypotheses from entering specification.
Structure a probe plan for high-risk assumptions by prioritizing blocking hypotheses, defining validation methods, setting sample sizes, establishing success or failure thresholds, and documenting evidence paths within a strict timebox.
You need a probe plan for risk management during product discovery and specification phases when fast, evidence-based decisions are required to decide whether to push mainline development forward or pivot.
Yes, you can use probe plans to prioritize product hypotheses by focusing specifically on blocking assumptions and those with high impact but low confidence, avoiding the trap of trying to validate too many hypotheses at once.
Small probes for product validation are limited to testing one or a few key assumptions at a time, meaning they are not suited for comprehensive feature testing but rather for quickly closing blocking hypotheses before entering specification.
autostay-krPrioritize product assumptions using an Impact x Risk matrix and suggest validation experiments.
MinhDuyDEVClarify ambiguous requirements through hypothesis-driven discovery interviews before implementation.
imbue-aiProcess managed agent lifecycle state events and trigger automated responses.
openyakPlan, conduct, and synthesize user research studies into actionable insights.
