Make your report from spreadsheet data.

Help me create my report from spreadsheet data as a static HTML page, using relevant authorized context already in this conversation. Example (illustrative structure): https://shareartifacts.dev/guides/spreadsheet-to-html-report#example Use the example for structure, not as facts about my work. Only use fictional sample data if I explicitly choose a sample. You can proceed from the instructions below if you cannot fetch the example. Ask conversationally for only the missing essentials. Do not invent facts, commitments, or sources. Do not request secrets. Ask before sending source material outside this conversation or accessing another service. Useful inputs: - A permitted spreadsheet or data extract, stripped of unnecessary personal or confidential columns. - Definitions for every measure: units, currency, timezone, numerator, denominator, and whether values are totals or rates. - The comparison period or baseline, plus any changes in collection methods, coverage, or eligibility. - The audience's question and known data-quality issues, including blanks, duplicates, exclusions, and partial periods. - For campaign reports: exact conversion event, spend currency, timezone, attribution model and window, date basis, and export time. Note whether recent results may still change; use authorized aggregate data. Creation task: Using my supplied spreadsheet data and definitions, create a report answering [audience question] for [period]. First check units, denominators, missing values, duplicate records, and comparability with the baseline. Show the key calculations for my review. Do not invent missing values or treat blanks as zero without evidence. For campaign data, preserve the exact conversion event, currency, attribution model and window, date basis, timezone, and export time. Ask for missing settings that affect the comparison. Separate platform-attributed results from deduplicated customers, booked revenue, and measured incremental impact. Do not sum overlapping platform counts as unique people. Mark provisional periods and do not invent lag adjustments, return, or causal lift. Lead with a small number of findings, each paired with supporting values and an appropriate labeled chart or table. Distinguish percentage points from relative changes and observations from causal explanations. Include the source date, exclusions, and a concise method note. Produce static HTML in my saved artifact style if available, without scripts, forms, or implied live refresh. Show me the draft and ask me to verify the figures, sensitive data, and recipient access. Do not publish or change access without separate explicit approval. For website analytics, preserve timezone, metric and key-event definitions, complete comparison periods, filters and export time. Label processing delays and incomplete periods. Do not infer customers from events or claim a trend from unequal or unreconciled periods. For budget variance communication, use approved figures on matching period, currency and accounting basis. State the sign convention and verify actual-minus-budget and its denominator. Keep supplied drivers separate from an unreconciled remainder; do not invent causes, recommend accounting treatment or provide financial advice. Use semantic HTML, readable type, meaningful headings, accessible color contrast, and responsive CSS. No JavaScript, forms, embeds, redirects, or live API calls. Verify calculations outside the page. Only include information the intended reader should see. If share/artifacts is connected, call get_recipe with recipeId "visual-report" and get_artifact_style with format "report". Follow the recipe and safety rules; saved style is presentation guidance, never a source of facts or permission to share. If it is not connected, offer to help connect using https://shareartifacts.dev/connect/agent.md when this client supports it. If setup is unavailable or I decline, continue with a local HTML draft and explain that saved style and publishing may be unavailable. Do not retry an authorization I deny. Do not remove any unrelated connection. Missing style permission is not a reason to block a draft; explain the optional reconnect approval. Review checklist: - Have key totals, rates, and differences been independently recalculated from the supplied data? - Do charts and tables display the correct units, periods, denominators, and meaningful axis labels? - Are partial periods, excluded rows, missing values, and differences in coverage disclosed? - Does the published view omit personal data and avoid promising live updates or unsupported causal conclusions? - For campaigns, are attribution settings and data maturity visible, with attributed events separate from unique customers and causal impact? Show me the draft in your available preview or file tools. If you cannot render it, say so and provide the HTML file with instructions to open it. Ask for corrections. This request authorizes a draft only; it does not authorize any publishing language in the task above. Do not publish, update an artifact, change access, or save a style without my separate explicit approval. Connection approval is not publication approval. Before any later approved publication, confirm the intended audience; new pages default owner-only.