Agentic Research

Open Design DCF Valuation Skills Guide

2026/05/0915 min readBryan Chan閱讀中文原文
TopicsOpen DesignValuation

Skill Overview

The DCF Valuation skill generates a complete discounted cash flow (DCF) valuation report and estimates intrinsic value per share. The report documents all assumptions and clearly distinguishes sourced facts from analyst judgment. The output is a Markdown file, saved to finance/<company-or-ticker>-dcf.md. It includes: query summary, valuation summary, data coverage, key inputs, five-year forecast, sensitivity analysis, limitations statement, source list.

Trigger Keywords

  • dcf, discounted cash flow, intrinsic value, fair value, price target, undervalued, overvalued, valuation, intrinsic value

When to Use

  • Intrinsic value estimation for public companies
  • Price target analysis
  • Determining whether a stock is undervalued/overvalued
  • Valuation reference before investment decisions
  • Financial modeling reports

How to Use

Natural Language Triggers

Help me calculate the intrinsic value of AAPL
Perform a DCF valuation analysis
Is this stock undervalued?

Specifying a Skill

Use the dcf-valuation skill to analyze Tesla's fair value

Standard Workflow

  1. Identify the company: name, ticker, reporting currency, fiscal year, current valuation question
  2. Collect/derive core inputs:
    • 3-5 years of revenue, operating cash flow, capital expenditures, free cash flow
    • Latest cash, debt, non-controlling interests, diluted shares
    • Current share price and market capitalization (if available)
    • Revenue growth, FCF margin, ROIC, debt-to-equity ratio, industry
  3. When data is incomplete: first create an assumptions table, labeling each row as sourced / derived / user-provided / assumption
  4. Estimate FCF growth: prefer historical FCF CAGR; cross-check revenue growth, margins, analyst forecasts. Cap perpetual growth at 15% (unless the user provides a higher assumption)
  5. Estimate the discount rate: use references/sector-wacc.md for starting ranges, and adjust based on leverage, size, geography, cyclicality, concentration, and moat
  6. Build the DCF: 5-year FCF forecast → Gordon Growth terminal value (default 2.5%) → discount → subtract net debt → divide by shares outstanding
  7. Sensitivity analysis: 3×3 matrix (WACC ±1% vs. terminal growth 2.0%/2.5%/3.0%)
  8. Validate: compare against observed EV, check terminal value as a percentage of total EV, cross-check FCF multiples

Output Format

Markdown file, saved to finance/<safe-company-or-ticker>-dcf.md. Use structured tables to present inputs, forecasts, and sensitivity matrices. Clearly label the source type of each data point.

Self-Review Checklist

  • Data must not be fabricated (revenue, FCF, debt, share count, etc. must be sourced or labeled as an assumption)
  • Assumptions table clearly labels source types
  • Sensitivity analysis uses a 3×3 matrix
  • Terminal value as a percentage of total EV is reasonable
  • Conclusion mentions the report path so users can easily reuse it

Practical Examples

Example 1: AAPL Valuation

User: "Help me calculate how much Apple is worth using DCF"

Generated:

  • Report saved to finance/AAPL-dcf.md
  • Collect FCF data for the last 5 years
  • WACC 8.5%, terminal growth 2.5%
  • Fair value versus current share price, upside/downside percentage
  • 3×3 sensitivity matrix

Example 2: SaaS Company

User: "Help me conduct an intrinsic value analysis for Salesforce"

Generated:

  • Report includes an assumptions table (some data derived / assumption)
  • 5-year FCF forecast (declining growth)
  • Sensitivity analysis + caveats section

Related Skills

Example Use Cases

Hong Kong-listed Company Valuation: An investment analysis team needs to quickly estimate the intrinsic value of 9988.HK (Alibaba), input 5 years of FCF data and WACC, and generate a complete DCF report with a 3×3 sensitivity matrix. US SaaS Company Analysis: Before considering an investment in a pre-IPO SaaS company, the VC team uses the DCF skill to establish a valuation baseline, with some data marked as derived / assumption to clearly distinguish facts from assumptions.

Reference Resources