china-equity-ic-rigor

China Equity IC-Rigor

This skill is the PM red-team layer on top of china-equity-research. Use the base skill to do the underlying multi-agent research; use this one to harden the deliverable to the standard a buy-side PM will sign off on.

The core insight encoded here: an institutional-grade equity opinion letter survives PM challenge not because it has a strong view, but because every specific number is sourced, every transformation reconciles, every headline acknowledges what it's contingent on, and every "what would reverse it" trigger has a numerical denominator. The bugs that kill a memo in IC are almost always mechanical (math doesn't multiply, definitions don't match, anchors aren't verified) — not directional.

When to use this skill

Trigger when ANY of the following appear, even in passing:

• Ticker (000XXX / 6XXXXX / H-share) + "投资意见书" / "IC memo" / "意见书"
• "Red team this", "score this memo", "PM review", "round N", "what would push this from 8.x to 9.x"
• Headline-language requests: 中位预期收益, 情景加权区间, 仓位建议, 减仓/加仓
• Explicit references to S1-S5 sources, 强多/多头/基础/空头/强空 scenarios, 三估值法 reconcile, GM taxonomy, bear bridge, what-would-reverse, A0 tail
• Multi-audience derivatives: 精简版 / IC pre-read / IC debate script / 零售版 / 非专业版
• The user is critiquing a memo and the language sounds like a PM ("the math doesn't add up", "where does this number come from", "this is hand-wavy", "you can't make that claim without an S1")

If the request is for initial fundamental research (no opinion letter framing yet), use china-equity-research directly. Add this skill once the memo construction or red-team phase begins.

Workflow

The work proceeds in five phases. Phases 0-3 produce the institutional version. Phase 4 hardens it. Phase 5 derives audience variants. Phases can interleave when the user explicitly directs it (e.g. "build the institutional and IC versions in parallel").

Phase 0 — Foundational research (delegate to china-equity-research)