Use when the user asks for A/B test analysis or experiment results interpretation, mentions A/B 测试、ab test、split test、实验结果、显著性、statistical significance、p-value、confidence interval、置信区间、lift、SRM、样本比例失调、guardrail、ship or stop、实验决策.
Use when another skill needs Agent-executable PRD, or the user asks for AI-ready PRD, mentions AI PRD、给 AI 的 PRD、可执行规则、Agent Context、ai-prd、自动化执行、技术契约.
Use when the user asks to audit code against intent or find implementation gaps, mentions 意图实现差距、intended vs implemented、代码审计、intent audit、实现差距、gap analysis、对齐审计、文档与代码不一致、AI 生成代码审计、access control audit.
Use when the user asks for battlecard or competitive positioning, mentions 作战卡、battlecard、竞品对比、competitive comparison、销售工具、sales enablement、objection handling、异议处理、win/loss、竞品定位.
Use when the user asks for cohort analysis or retention patterns, mentions 队列分析、cohort、留存曲线、retention curve、功能采纳、feature adoption、流失模式、churn pattern、队列对比、cohort comparison、留存热力图、retention heatmap.
Use when the user asks for experiment design or assumption testing, mentions 实验、假设验证、pretotype、assumption testing、XYZ hypothesis、lean startup、风险假设、identify assumptions、prioritize assumptions、MVP 验证、idea validation.
Use when the user asks for go-to-market or launch strategy, mentions GTM、走向市场、beachhead、go-to-market.
Use when the user asks to handoff or continue in another session, mentions 交接、handoff、切换会话、continue later、session handoff、换 agent、接续工作、上下文压缩、compact session、知识转移、knowledge transfer、条件触发、condition-based trigger.