How to use the consulting-analysis agent skill (research reports)

What is an agent skill, and what does this one make the agent do?

An agent skill is a folder containing a SKILL.md file with instructions that an AI agent loads when a request matches the skill's description. This skill is called consulting-analysis and its description covers requests to generate, create or write professional research reports, including market analysis, consumer insights, brand analysis, financial analysis, industry research, competitive intelligence, investment due diligence and similar analytical reports.

The skill works in two phases. In phase 1 the agent drafts an analysis blueprint: a chapter outline, the data each chapter needs and the analytical frameworks to apply, such as SWOT, Porter's Five Forces or PESTEL. In phase 2, after data has been gathered by other tools or skills, the agent writes a Markdown report in a consulting tone, with charts, comparison tables and inline citations.

• Phase 1: chapter skeleton, data query requirements and analysis logic.
• Phase 2: a final report built from the data you supply, with embedded charts and strategic insights.
• Citations are formatted to GB/T 7714-2015 by default, and the report language follows an output_locale setting (default zh_CN, Chinese).

Install the consulting-analysis skill

Download the skill ZIP from this page, then unzip it into one of two locations. For a personal setup, unzip it into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder. The skill's package contains a SKILL.md file.

If you use the Claude apps instead, open Settings, go to Capabilities, then Skills, and upload the ZIP there.

• Personal: ~/.claude/skills/
• Project only: .claude/skills/ inside the project folder
• Claude apps: Settings → Capabilities → Skills, then upload the ZIP

How to trigger the skill with a natural request

You do not need a special command. The skill loads when your request matches its description, so ask for the kind of report you want in plain words. Give the research subject first, and add scope if you have it, such as geography, time range, industry segment or target audience.

For phase 1, ask for a framework before any data exists. For phase 2, paste or attach the data summaries, framework or chart files you already have and ask for the report to be written from them.

• "Draft an analysis framework for a market analysis of Gen-Z skincare in China."
• "Plan a competitive intelligence report on the electric vehicle industry, focusing on SWOT and Porter's Five Forces."
• "Write a consulting-style brand analysis report from the data summary I've attached."

What the skill needs and what it does not do

The listed requirements are nothing, the skill runs no scripts and it does not use the network. The SKILL.md itself does not collect data. Data must come from you or from other skills or tools you run alongside it.

The skill is instructed not to invent numbers. If data is missing, the report should say data is not available rather than fabricate figures. Even so, the skill only sees the data you supply, so every number should be checked against its original source before you rely on it.

• No listed requirements, no scripts run, no network use.
• Data gathering is not automated by this skill.
• Output language and citation style are set for Chinese (zh_CN) and GB/T 7714-2015; changing them needs adjustment.

Safety notes for scripts and network access

According to its listing, this skill runs no scripts and uses no network access. The package lists a single file, SKILL.md. Because the skill does not gather data itself, any web search or data collection happens only through other tools or skills you have installed separately, and you should review those on their own terms.

Treat the output as a draft. The skill's instructions do not guarantee accuracy, so check each figure, chart and citation against its source before sharing the report.

• The skill's files: SKILL.md only.
• Check every figure against the original source before publishing.
• Review any other skills you run alongside it, since they handle data collection.

Tips for better reports

Run phase 1 first. A clear chapter outline and data requirement list makes it easier to know what to collect before you ask for the report.

Supply clean data summaries with sources noted. The report can only be as traceable as the input you give it.

Name the frameworks you want, or let the agent choose based on the domain. Common choices include SWOT for brand work, Porter's Five Forces for industry structure and PESTEL for macro scans.

• Start with phase 1 to get the chapter plan.
• Give scope: geography, time range and audience.
• Name the frameworks you need.
• Keep sources with every data point.