How to use the Valuation Model Methods agent skill

What an agent skill is and what this one makes the agent do

An agent skill is a folder containing a SKILL.md file with instructions the agent reads when a request matches the skill's description. This skill is a single file, SKILL.md, and it contains no code to run.

Once installed, when you ask the agent to value a company, it follows the methods defined in the skill rather than improvising. The skill is described as a systematic framework for first-pass valuation of China A-share companies.

• Absolute valuation: DCF (with WACC and terminal value), DDM (two-stage and Gordon models) and SOTP (sum of the parts).
• Relative valuation: PE band (historical PE percentiles), PB-ROE matrix and EV/EBITDA.
• Sensitivity analysis, such as a WACC versus perpetual-growth grid.
• Valuation-trap detection, including one-off gains, goodwill risk, share dilution and receivables quality.
• Cross-validation: use at least two methods and check whether they differ by more than 30%.

How to install the valuation-model skill

Download the ZIP file for the skill from this page. Then choose one of the two installation routes below.

For a personal install that applies to all your projects, unzip the file into ~/.claude/skills/. For a single project, unzip it into that project's .claude/skills/ folder. Either way, the skill folder should contain SKILL.md.

If you use the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.

How to trigger the skill with natural requests

You do not need a special command. Describe the task in plain language that matches the skill's purpose, naming the company and, if you have them, the figures. The skill's description covers valuing a company, DCF, DDM, SOTP, PE band, PB-ROE, EV/EBITDA, sensitivity analysis and valuation traps.

Including the company name or stock code and the method you want helps the agent pick the right section of the skill.

• Value Kweichow Moutai using a DCF and a PE band, and show a sensitivity table.
• Do a first-pass valuation of this company using at least two methods and flag any valuation traps.
• Is this stock a valuation trap? Check one-off gains, goodwill and share dilution in the figures I paste.
• Which valuation method fits a high-dividend bank? Run a DDM.

What you get back: an example output

The skill defines an output format with a valuation summary table, a method-by-method weighting, a sensitivity matrix and a checklist of trap checks. The figures in the skill are illustrative examples, not current market data.

For example, the output template combines a DCF value, a comparable-PE value and a PB-ROE value into a weighted composite target price, then compares it with the current price to show implied upside.

• Valuation summary: each method with its per-share value, weight and notes.
• Sensitivity analysis: WACC against growth rate.
• Valuation-trap check: a tick-box list, such as whether goodwill is above 30% of net assets.

What the skill needs and safety notes

The skill lists no required tools or packages. It does not run scripts and does not use the network, so it works purely from the information you give the agent and the reference ranges in the file.

Because the skill makes no network calls, it cannot fetch current prices, financial statements or PE history. You need to supply those figures yourself, or have the agent use other tools you have enabled.

The reference WACC, growth, multiple and PE ranges are generic and calibrated to China A-shares. Check them against current data before using any result.

• Requirements: nothing listed.
• Scripts: none.
• Network access: none.
• Data: you provide the company's financials and price; the skill does not supply market data.

Tips for getting better valuations

Give the agent clean inputs. The DCF needs a forecast of revenue, EBIT, capex and working capital; the PE band needs a history of PE_TTM values. Missing inputs will weaken every method.

Ask for at least two methods and compare them. The skill itself recommends a cross-check and says to revisit assumptions when methods differ by more than 30%.

Treat the sensitivity table as the real answer. A single value depends heavily on WACC and terminal growth; the grid shows the range.

• Use DDM only for companies with a long dividend record and stable payout.
• For cyclical companies, the skill advises PB and EV/EBITDA rather than PE.
• For loss-making companies, the skill points to price-to-sales and EV/Sales.
• Ask the agent to run the trap checklist separately on the reported figures.