Guide · Agent skill
How to use the Market Sizing (TAM, SAM, SOM) agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The market-sizing skill turns your AI agent into a market analyst that estimates the total market (TAM), the part your product could realistically serve (SAM), and the share you could plausibly win (SOM) for any product or opportunity you name. It is written for founders and product managers preparing pitches or business cases, students learning these concepts, and teams weighing a new market or region. After this guide you will know what the skill does, how to install it, how to phrase requests so it triggers, and where its output should not be trusted without checking.
Market Sizing (TAM, SAM, SOM)phuryn/pm-skills · MIT Download the skill Quick start
- 1Download the ZIP. Download the market-sizing skill ZIP from this page.
- 2Unzip into your skills folder. For Claude Code, unzip into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder for one project.
- 3Or upload in the Claude apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP.
- 4Write a specific request. Name the product, the geography or segment, and the purpose, such as an investor pitch or market entry decision.
- 5Provide your own data. Paste pricing, customer counts, reports or competitor information so the bottom-up estimate uses real inputs.
- 6Check the assumptions. Review the confidence ratings and sources, and validate the most uncertain assumption before relying on the result.
What is an agent skill?
An agent skill is a folder of instructions that an AI agent can load when a task matches it. The core file is SKILL.md, which holds a short description of when the skill applies and the detailed instructions the agent follows.
This skill is published by phuryn as part of the pm-skills collection, under the pm-market-research category, and is released under the MIT licence.
What the market-sizing skill makes the agent do
Once loaded, the agent follows a set of analysis steps: define the market boundaries, estimate top-down from total industry size, estimate bottom-up from customers times price times frequency, scope the serviceable part (SAM), estimate the obtainable share over 1-3 years (SOM), project growth over 2-3 years, and map the key assumptions.
The output is organised into market definition, TAM, SAM, SOM, a summary table of current estimates and projections, growth drivers, and key assumptions with a confidence level for each. The skill asks for both top-down and bottom-up figures so they can be cross-checked against each other. It can also search the web for current industry data when you ask it to.
- Market definition: problem space, customer segments, geography and constraints
- TAM, SAM and SOM, each with reasoning and percentages of the level above
- Top-down and bottom-up estimates, reconciled against each other
- Growth projections and a summary table
- Assumptions rated high, medium or low confidence, with ways to validate them
How do you install the skill?
There are two routes. For Claude Code, download the ZIP from this page and unzip it into your personal skills folder at ~/.claude/skills/, or into a project's .claude/skills/ folder if you want it only for that project. For the Claude apps, upload the ZIP under Settings → Capabilities → Skills.
Your agent must be able to load SKILL.md skills. If the skill does not appear or trigger, check that the unzipped folder contains SKILL.md at the expected level for your setup.
- Claude Code, personal: unzip into ~/.claude/skills/
- Claude Code, one project: unzip into .claude/skills/ inside the project
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
~/.claude/skills/How do you trigger the skill with a natural request?
The skill's description tells the agent to use it when sizing a market opportunity, estimating addressable market, preparing for investor pitches, or evaluating market entry. Ask in those terms and name the product, and include any constraints such as geography, industry vertical or customer type.
The skill's instructions use the text you give after the skill name as the market to estimate, so a specific request gives a more useful result than a vague one.
- Size the market for a meal-planning app for UK families, including TAM, SAM and SOM
- Estimate the addressable market for our invoicing tool for freelance designers in Germany
- I'm preparing an investor pitch. Work out TAM, SAM and SOM for a B2B expense tool for mid-sized law firms
- Should we enter the Southeast Asian market for electric scooter rentals? Evaluate the market size first
What should you provide to get a better estimate?
The skill reads any market research, industry reports, financial data or competitor information you paste or attach, and analyses them directly. Providing your own numbers, such as your price point, target customer count or current traction, lets the agent build the bottom-up estimate and the SOM on something real rather than assumptions.
Be explicit about the boundaries. A market defined as 'UK households with children under twelve who cook at home' will produce a more usable SAM than 'food apps'.
- Your pricing and how often customers buy
- The customer segments and geography you care about
- Any industry reports or competitor data you already have
- Your current traction, sales capacity and channels
What are the safety notes: scripts and network use?
This skill does not run scripts. Its only file is SKILL.md. It does use the network: the instructions tell the agent to use web search for current market data, industry reports and benchmarks, so your agent needs web search access for that part. The skill states that web search is recommended but is not a hard requirement, so it still works from the material you supply and its general knowledge.
Because the agent may search the web, figures it finds can be outdated or come from weak sources. The skill itself asks the agent to cite sources and label estimates versus data, so check those citations before you rely on them.
- Scripts: none. The package contains SKILL.md only.
- Network: yes, for web search when enabled and when you ask for current data
- Web search is recommended, not required
What are the limits and tips for good results?
The skill is not a precise financial forecast. Its output is a structured estimate built on assumptions and web sources, so verify figures before using them in filings or investment decisions. For markets with little public data, the agent may fill gaps with plausible but unsupported numbers, so treat any figure without a source as a guess.
Use the confidence ratings and the 'how to validate' section to decide which assumptions to test first, for example by interviewing customers or checking an official statistics source.
- Ask for both top-down and bottom-up figures and compare them; a large gap means a weak assumption somewhere
- Check each cited source and its date before using the number
- Ask the agent to recommend data sources for your most uncertain assumption
- Say whether you want value (revenue) or volume (users or units) sizing
- For international markets, ask about currency and purchasing power
What people use it for
- Sizing the addressable market for a new B2B software product before writing an investor deck
- Comparing two candidate markets or regions to decide where to launch first
- Building the market section of a product business case with top-down and bottom-up figures
- Teaching students how TAM, SAM and SOM differ and how each is estimated
- Stress-testing a founder's market claim by checking whether the bottom-up and top-down numbers agree
Frequently asked questions
Is the market sizing agent skill free?
The skill is released under the MIT licence, which governs use of its code and files. Check the licence text on the source page for the full terms. You still need an AI agent that can load SKILL.md skills to use it.
How do I install the market-sizing skill?
Download the ZIP from this page. In Claude Code, unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder. In the Claude apps, upload the ZIP under Settings → Capabilities → Skills.
What does the agent need to run this skill?
It needs an AI agent that can load SKILL.md skills. Web search access is recommended so it can find current market data, but the skill says it is not a hard requirement. The skill does not run scripts, but it does use the network for web search.
How accurate are the TAM, SAM and SOM numbers?
They are structured estimates built on assumptions and web sources, which can be outdated or inaccurate. Verify figures before using them in filings or investment decisions. The skill rates each assumption as high, medium or low confidence to show where to be careful.
Why does the skill use both top-down and bottom-up estimates?
Top-down starts from total industry size and narrows it, while bottom-up builds from customers times price times frequency. Calculating both lets the agent cross-check them and reconcile the difference, which makes the estimate easier to defend.
Will the skill work for a market with little public data?
It can be unreliable there. The skill's own limits note that the agent may fill gaps with plausible-sounding but unsupported numbers. Supply your own data, and treat any figure without a source as a guess to test.
How is this different from a generic market size prompt?
The skill gives the agent a fixed method and output structure: market definition, both estimation approaches, SAM and SOM scoping, growth projections, and an assumptions table with confidence levels. A plain prompt often skips the reconciliation and assumption checks.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you are preparing an investor pitch, learning how TAM, SAM and SOM are estimated, or deciding whether to enter a market, and you will verify the sources it cites. Pick something else if you need precise financial forecasting, or if the market has little public data, since the agent may fill gaps with unsupported numbers.
Download the skill