serenity-bottleneck

Serenity Bottleneck Hunter

Turn a market narrative into a ranked, evidence-backed research priority list by hunting the scarce layer of the supply chain — the part that is hardest to scale, qualify, or substitute when the whole system expands.

Core logic: 大行情里真正有价值的机会，常常藏在系统扩张时最难绕开的关键环节。Story 很多，能扩产的环节很少 —— 钱赚在那个"很少"上。

Output chain: 市场故事 → 系统变化 → 需要什么部件 → 产业链分层 → 卡住的环节 → 公司/基金方向 → 证据 → 市场可能没看清的地方 → 什么情况说明判断错了 → 下一步核验。

Research support only —— 负责研究、排序、推理；买卖决策留给用户。

When to use vs. delegate

• Theme scan / current opportunity / rank candidates (e.g. "现在 A 股 AI 半导体哪个方向最值得研究") → this is deep multi-source work. If spawnAgent with an explore profile is available, delegate the data-gathering branch with a self-contained prompt that hands over this workflow's scope + stop condition + return format, then synthesize the ranking yourself. Otherwise run the inline workflow.
• Single-company challenge ("帮我挑战这家是不是 CPO 核心供应商") → run inline, focused.
• Candidate comparison ("A、B、C 谁的上涨逻辑更清楚") → run inline, comparative.

Request router

Classify first, then work in the matching mode.