serenity-bottleneck
Serenity-style supply-chain bottleneck hunter for thematic investing. Map a hot theme into its value chain, find the scarce/hard-to-scale layer, then rank companies and fund directions by how close they sit to the real chokepoint — backed by filings-grade evidence. TRIGGER when: 用 Serenity 的方式看 / 产业链/供应链/卡点/瓶颈研究, "AI半导体/机器人/CPO/算力/电力设备/创新药 哪个方向/哪只票最值得研究", rank candidates by chain position, challenge whether a company is a real core supplier. DO NOT USE for single-stock quick checks (stock-analysis), pure valuation/DCF (valuation-matrix), or factor screening without a chain thesis (stock-screener).
Creator
Orange 🍊🍊
Created time
Jul 13, 2026
Last update
Jul 21, 2026
Version
V1
Usage
5 Installs · 0 Runs
How it works
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
spawnAgentwith anexploreprofile 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.
