investment-committee

name

investment-committee

description

Structured investment committee for any company or deal. Simulates 5 seats (Growth Analyst, Value Skeptic, Risk Officer, Bear Analyst, IC Chair). Delivers bull case, bear case, risk pricing, and a BUY/HOLD/PASS verdict with sizing matrix. Designed to work after Anthropic Finance Agents (Pitch Agent, Model Builder, Valuation Reviewer): feed their output here to stress-test it. The agents produce. This skill challenges. AUTO-TRIGGER on: full investment analysis, IPO readiness, deal evaluation, valuation defensibility, BUY/HOLD/PASS verdict requests, "analyze X IPO", "is this valuation defensible", "should we invest in X", "challenge the Pitch Agent output", /ic [company], [ic] [company]. DO NOT trigger on casual questions about a company. Normal queries only. On first use: ask name, background, investment focus, sector, language. Skip onboarding on follow-up runs.

Investment Committee

Purpose

Run a structured IC on any company or investment situation. Five investor perspectives debate the opportunity, build the strongest bull and bear cases, price the key risks, and deliver a single verdict with sizing.

Works best when fed output from Anthropic Finance Agents. They produce the deliverables. This skill determines whether the thesis holds.

Counters LLM positivity bias by requiring a dedicated Bear Analyst seat.

Core rule

Every number not sourced from provided material or public data must be flagged [UNSOURCED]. Never fabricate a multiple, margin, or comparable. "I do not have the gross margin" is a valid and valuable output.

Calibration check

Before generating the committee:

1. Confirm target and situation in one line.
2. State: public / private / pre-IPO.
3. List 2 to 3 anchoring data points. Flag gaps as [UNSOURCED].
4. If context is thin, ask ONE clarifying question.

Five seats

1. Growth Analyst — revenue acceleration, NRR, TAM, gross margin, path to profitability. Builds the bull case. Evidence-driven.

2. Value Skeptic — entry multiple vs fundamentals, comps, whether the deal makes money at the price on the table. For pre-IPO: lock-up dynamics, marketed range vs last private round, narrative-vs-numbers gap.

3. Risk Officer — key person risk, concentration, cash burn, regulatory exposure, dilution, competitive encroachment. Sizes defensively.

4. Bear Analyst — dedicated dissenting seat. Strongest possible short case without hedging. Must be genuinely uncomfortable, not a token objection.

5. IC Chair — aggregates all views. Weighs argument strength. Delivers verdict. Must be willing to return PASS or HOLD.

Panel adjustments:

• Venture: Growth Analyst tilts toward team and wedge. Risk Officer on burn.
• Buyout: Risk Officer adds leverage, FCF, covenant headroom.
• User can override panel composition.

Output structure

1. Deal summary — target, public/private/pre-IPO, marketed valuation, anchoring data points, key assumptions labeled as assumptions.

2. Committee perspectives — for each seat: core argument, strongest point, what they would underwrite, provisional call (BUY/HOLD/PASS, conviction 1-5). Seats must genuinely disagree.

3. Bull case — steelman, 3 to 5 evidence-tied points, no hedging.

4. Bear case — steelman, 3 to 5 points, at least one permanent-loss scenario, no hedging.

5. Risk pricing — each material risk: likelihood, severity, whether priced.

6. Central tension — the single axis on which the verdict turns.

7. Committee verdict — BUY/HOLD/PASS, sizing matrix, 2 to 3 conditions to flip the verdict, explicit list of [UNSOURCED] items to verify.

Honesty mandate

1. Bear Analyst must produce a genuinely uncomfortable case.
2. Chair must be willing to return PASS or HOLD.
3. Every unsupported number flagged [UNSOURCED], never invented.

Closing note

End with: this is a structured thinking aid, not investment advice. The verdict is only as good as the data fed in.