Guide · Agent skill
How to use the Market Research Reports agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The Market Research Reports skill is a set of instructions, templates and optional Python scripts that teaches an AI agent to build market research with traceable sources, traceable calculations and market size estimates shown as scenarios. It is aimed at analysts, marketers and students who need numbers they can defend. By the end of this guide you will know what the skill changes about the agent's output, how to install it, how to ask for it and what to check before relying on the result.
Market Research ReportsK-Dense-AI/scientific-agent-skills · MIT Download the skill Quick start
- 1Download and unzip. Download the skill ZIP from the directory page and unzip it into a folder.
- 2Place the folder. Move the skill folder into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder.
- 3Or upload in the apps. In the Claude apps, open Settings, go to Capabilities, then Skills, and upload the skill.
- 4State the decision and scope. Describe the decision, market, geography, period and currency so the agent can fix the market definition.
- 5Provide your sources. Give the agent the data you have access to, since it cannot fetch paid or live data on its own.
- 6Run the validation scripts. Run the ledger validation, citation audit and sizing calculator with Python 3.11+ to check the output.
What an agent skill is and what this one changes
An agent skill is a folder of instructions and supporting files that an AI agent loads when a request matches the skill's description. This skill is published by K-Dense-AI in the scientific-agent-skills repository, under the MIT licence. Its description covers market definition, industry and customer evidence, competitive landscapes, TAM/SAM/SOM reconciliation, forecast sensitivity and auditable report scaffolds.
Once installed, the agent is steered to link each factual claim to a numbered source ID, keep calculations traceable to their inputs, and show TAM, SAM and SOM as scenarios with stated assumptions rather than one number. It also points to templates for a source ledger, a claims ledger, a competitor feature matrix and forecast sensitivity.
- Every source gets a stable ID such as S-001, and every claim gets an ID such as C-001.
- Market size is built as at least two scenarios, not a single figure.
- The agent is told not to invent citations, quotes or market shares, and not to present estimates as certain.
- Missing public feature evidence is recorded as unknown, not as no.
Install the skill from the ZIP
Download the ZIP for this skill from the directory page and unzip it. For a personal install, place the skill folder in ~/.claude/skills/. For a single project, place it in that project's .claude/skills/ folder.
If you use the Claude apps instead, open Settings, go to Capabilities, then Skills, and upload the skill there.
- Personal install: ~/.claude/skills/ (available across your projects).
- Project install: .claude/skills/ inside the project folder (available only to that project).
- Claude apps: Settings → Capabilities → Skills, then upload.
Trigger the skill with natural requests
You do not need a special command. The skill matches on its description, so ask for the kind of work it covers in plain language. Name the decision, the market and the output you want, and say whether you need sources cited.
The agent will ask a focused question only when a missing choice would change the result, such as the geography or the currency. Otherwise it states a provisional scope and continues.
- "Size the market for ergonomic office chairs in Germany for a 2027 board decision, with every number sourced."
- "Build a competitor feature matrix for three payroll software vendors, marking anything I can't verify as unknown."
- "Reconcile our TAM and SAM figures and show the downside, base and upside scenarios."
- "Set up a source ledger and claims ledger for this market report draft."
Use the templates and optional validation scripts
The skill ships templates for each part of the audit trail: a source ledger, a claims ledger, a consistency check, a market sizing scenarios file, a forecast sensitivity file and a report manifest. It also includes a LaTeX-friendly formatting guide and reference notes on evidence and methods.
Two offline scripts check your work. Run them from the skill folder with Python 3.11 or later. The first validates a source ledger CSV; the second audits claim-to-source mappings. The third computes market sizing from a JSON file of scenarios.
- Source ledger template: assets/source_ledger_template.csv
- Claims ledger template: assets/claims_ledger_template.csv
- Market sizing template: assets/market_sizing_scenarios_template.json
python3 scripts/validate_evidence_ledger.py data/source_ledger.csv
python3 scripts/audit_claim_citations.py data/claims.csv data/source_ledger.csv
python3 scripts/calculate_market_sizing.py assets/market_sizing_scenarios_template.jsonSafety notes: scripts, network and data access
The bundled scripts make no network, LLM or image calls, and the skill lists Python 3.11+ standard library as its only requirement for them. The optional LaTeX template needs XeLaTeX or LuaLaTeX; Markdown, HTML or DOCX output is equally acceptable.
Online research is different. It needs network access you approve, and any source's own terms apply. The skill cannot verify sources itself, and it does not give you access to paid databases. It also says never to put an API key in a report or a bundled script.
- Scripts: run locally, standard library only, no network calls.
- Online research: only when you choose it, and only within each source's terms.
- Paid data: available only if you already have lawful access; the skill does not provide it.
- Do not paste confidential material or personal data into the report; the skill forbids PII disclosure and trade-secret acquisition.
Tips for getting a defensible result
Give the agent the measurement definition up front: product scope, buyer type, geography, period, currency and base year. Two estimates that do not share these are not directly comparable, and the skill will flag that.
Treat the output as a structured draft. The skill helps you organise evidence and expose uncertainty, but the data gathering and the judgement remain yours.
- Supply your own source files and let the agent map claims to them.
- Ask for sensitivity runs so you can see which assumption moves the answer most.
- Do not add manufacturer revenue to distributor or end-customer spend; the skill warns about this double counting.
- Keep the source ledger current, since publication and retrieval dates matter for revisions.
What people use it for
- Sizing a market for a manager or client with TAM, SAM and SOM shown as downside, base and upside scenarios.
- Building a competitor feature matrix where unverified features are marked unknown.
- Writing a student market analysis in which every number is tied to a numbered source.
- Reconciling top-down and bottom-up market estimates and listing the unresolved gap.
- Creating a reproducible audit trail of assumptions for a forecast that a team will revisit.
Frequently asked questions
Is the Market Research Reports skill free to use?
The skill is published under the MIT licence, as stated in its SKILL.md. The bundled scripts use only the Python standard library. Any paid data sources you use are not included and must be accessed under their own terms.
How do I install an agent skill like this one?
Download the ZIP from the directory page and unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder. In the Claude apps, you can instead upload it under Settings → Capabilities → Skills.
What does the skill require?
It needs Python 3.11 or later, using only the standard library, for the optional validation scripts. The LaTeX template needs XeLaTeX or LuaLaTeX if you want that format. Network access is needed only if you choose online research, and then source-specific terms apply.
Does the skill access live or paid market data automatically?
No. The skill states that it cannot verify sources itself and that the bundled scripts make no network calls. You need your own approved access to any database or live source you want to use.
Can it replace a market research firm or an investment advisor?
No. The skill does not replace doing the data gathering, and it excludes investment, legal, antitrust, tax, accounting and regulatory advice. Use it to structure and audit your analysis, and take those decisions to a qualified professional.
How is this different from just asking an AI to write a market report?
The skill adds structure: stable source and claim IDs, separated statement types, scenario-based sizing and validation scripts. A general request often yields a fluent narrative with unsourced numbers, which this skill is set up to avoid.
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 need a market size estimate you can defend, or a student or team needs every number traced to a source and assumption. Skip it for quick opinion pieces or one-off summaries where full sourcing is overkill, and do not rely on it for investment, legal, tax or regulatory decisions.
Download the skill