How to use the Build Analytical Report agent skill (build-report)

What is the build-report skill and what does it make the agent do?

An agent skill is a folder of instructions that an AI agent loads when a task matches the skill's description. This skill applies when you need a report that will outlast the chat, not a quick answer. The agent runs the analysis first, then uses this workflow to structure the reader-facing document.

The skill says the report should open with the answer, state the magnitude and the decision it implies, and only then cover methodology. Findings must be backed by evidence, and the agent is asked to separate verified facts from likely explanations, assumptions and open questions.

• Answer-first executive summary with decision implication
• Key findings with charts or tables only where they support a finding
• Recommendations tied to evidence, expected impact and an owner or next action
• Risks and limitations section covering data gaps and interpretation limits
• Sources and methodology section covering definitions, scope and source details

Which delivery format does the agent pick?

If you name a format, the agent honours it. Otherwise the skill defaults to Markdown, the most portable option. Self-contained plain HTML is used when a browser-readable report is useful and the host can render or inspect it. PDF, DOCX, slides or other formats are used only when you request them or the destination clearly requires them, and the host has suitable tools.

• Markdown (default)
• Self-contained plain HTML for a browser-readable report
• PDF, DOCX, or slides when requested or clearly required

How do you install the build-report skill?

You can install the skill in either of two ways. For a personal setup, download the ZIP from this page and unzip it into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder instead.

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

• Personal: unzip into ~/.claude/skills/
• Per project: unzip into .claude/skills/ inside the project folder
• Claude apps: Settings → Capabilities → Skills, then upload

How do you trigger the skill with natural requests?

You do not need a special command. Ask for a durable deliverable in plain words, and name the format and audience if you have them. The skill is meant for cases where a manager or stakeholder needs to act on the result.

If you only want a quick answer, say so. The skill itself says that when no file is wanted, the agent should give the same answer-first structure concisely in chat.

• "Turn these query results into a report my manager can act on."
• "Write an analytical report on last quarter's churn as an HTML file for the leadership team."
• "Give me a short inline brief on this, no file needed."

What does a good result look like?

Before handing back the file, the skill asks the agent to recalculate headline values and comparisons from the reviewed evidence, check that every chart matches its source data, and check titles, labels, units, denominators, date ranges and sorting. For HTML, PDF, DOCX or slides, the agent should render or inspect the actual artifact when the host supports it.

The agent is also told to keep credentials, local temporary paths and placeholder provenance out of the report. It returns the primary artifact path plus only the supporting notebook, SQL, data or chart files that are relevant.

• Headline numbers recalculated from the reviewed evidence
• Chart values checked against source data
• Units, date ranges and sorting checked
• Rendered artifact inspected when the host supports it

What do the optional helpers need, and what do they touch?

The core skill is instructions only. The package includes companion helpers for converting reports. The report-to-Google-Doc helper converts an HTML report into a DOCX for upload to Google Drive. It needs Python with python-docx installed, and it needs a Google Drive connector in the host agent for the upload. It does not call Google APIs itself.

The package does include scripts, so the listing states that it runs scripts. It does not use the network, according to the listing. There are also companion skills for Google Slides and PDF conversion, which each start from an existing HTML report and are invoked only when you explicitly ask for that output.

• Runs scripts: yes (for the DOCX/Google Doc helper and related files)
• Uses network: no
• Python with python-docx: needed only for the DOCX/Google Doc helper
• Google Drive connector: needed only for Google Doc upload
• The helper does not call Google APIs itself