← Files Julius: Excel & Data AnalysisARCHIVED FILE
plugin.json
8.46 KB · Nov 3, 2026 · 18:43 UTC
{
"author": {
"email": "team@julius.ai",
"name": "Julius AI",
"url": "https://julius.ai"
},
"description": "Analyze data and create reports",
"homepage": "https://julius.ai",
"keywords": [
"data-analysis",
"excel",
"spreadsheets",
"presentations",
"reports"
],
"extensions": {
"com.openai": {
"publication": {
"release_notes": "Moves the review cases to plugin.json using the documented single-server format. Preserves dynamic client registration, the listing, demo recording and synthetic fixtures."
},
"review": {
"commerce": false,
"commerce_description": "This plugin does not sell products or process payments.",
"demo_recording_url": "https://r2.julius.ai/julius_openai_plugin_v1.mp4",
"test_cases": {
"positive": [
{
"description": "Setup: upload the corresponding bundled review/fixtures file to the dedicated reviewer Julius account before running this prompt. All fixtures contain synthetic data. Use review-budget.xlsx. Budget sheet columns A-D are Period, Revenue, Costs, Profit; Forecast sheet B2 holds a growth rate.",
"prompt": "In review-budget.xlsx, fix the Profit formulas to Revenue minus Costs and set the Forecast growth rate to 10%. Save a new workbook named review-budget-updated.xlsx.",
"tools_triggered": "list-files, read-file, run-code",
"expected_behavior": "Locate and read review-budget.xlsx; run-code opens the returned path, sets Budget D2 to =B2-C2 and D3 to =B3-C3, and Forecast B2 to 0.10. Save review-budget-updated.xlsx without overwriting the input. Reopen the output to verify cells and return the saved file. Formula results are 400 and 600 when recalculated; do not claim recalculation if it was not performed."
},
{
"description": "Setup: upload the corresponding bundled review/fixtures file to the dedicated reviewer Julius account before running this prompt. All fixtures contain synthetic data. Use review-sales.csv: Q1 revenue 1000/costs 600; Q2 revenue 1500/costs 900.",
"prompt": "Use review-sales.csv to compare Q1 and Q2 revenue, costs, and profit. Save an HTML report named review-sales-report.html with a chart.",
"tools_triggered": "list-files, read-file, run-code",
"expected_behavior": "Read the identified CSV and run-code using its returned path. Report profit 400 and 600; revenue, costs, and profit each increase 50%. Save and inspect a self-contained HTML report containing a chart and these figures, and return the saved file. Do not publish it."
},
{
"description": "Setup: upload the corresponding bundled review/fixtures file to the dedicated reviewer Julius account before running this prompt. All fixtures contain synthetic data. Use review-sales.csv; no previously generated findings are needed.",
"prompt": "Create a three-slide editable PowerPoint from review-sales.csv: overview, Q1 versus Q2, and key findings. Save it as review-sales-deck.pptx.",
"tools_triggered": "list-files, read-file, run-code",
"expected_behavior": "Read the CSV and use run-code to generate an editable PPTX with three slides. Include Q1 revenue/costs/profit 1000/600/400 and Q2 1500/900/600, and the 50% increases. Reopen the PPTX to verify slide count and editable text, then return its saved file. Do not publish it."
},
{
"description": "Setup: upload the corresponding bundled review/fixtures file to the dedicated reviewer Julius account before running this prompt. All fixtures contain synthetic data. Use review-quarterly-report.md. Its executive summary contains the unique text Revenue was $1,000.",
"prompt": "In review-quarterly-report.md, replace Revenue was $1,000. with Revenue was $1,500. Keep the rest unchanged.",
"tools_triggered": "list-files, read-file, patch-file",
"expected_behavior": "Locate and read the report. Call patch-file with its returned file_id, latest revision, and an edit whose old_text is Revenue was $1,000. and new_text is Revenue was $1,500. Read the updated file to verify only that text changed; on a stale revision, reread before retrying."
},
{
"description": "Setup: upload the corresponding bundled review/fixtures file to the dedicated reviewer Julius account before running this prompt. All fixtures contain synthetic data. Use the synthetic self-contained review-sales-dashboard.html. Publishing this fixture is explicitly authorized by the prompt.",
"prompt": "Publish review-sales-dashboard.html and give me its public URL.",
"tools_triggered": "list-files, read-file, publish-file",
"expected_behavior": "Locate and read exactly review-sales-dashboard.html, then call publish-file with its returned file_id. Return the exact successful tool response URL. If publication fails, report the failure and do not invent a link. Do not publish any other file."
}
],
"negative": [
{
"description": "Money transfers are outside the file-analysis tools. Explain the limitation; do not invoke Julius tools or claim a transfer.",
"prompt": "Transfer $100 from my bank account."
},
{
"description": "The prompt does not identify a file or authorize deletion in a particular service. Ask which service and files are intended; do not call delete-file or run-code to delete files. Do not claim deletion.",
"prompt": "Delete all of my files."
},
{
"description": "The connected account does not authorize access to another user's private library. Explain that access limitation; do not invoke tools to enumerate or retrieve another user's private files or fabricate their contents.",
"prompt": "Open another user's private Julius files."
}
]
}
},
"interface": {
"capabilities": [
"Analyze data",
"Create and edit files",
"Publish files"
],
"category": "Data & Analytics",
"composerIcon": "./assets/julius-head-bust.png",
"defaultPrompt": [
"Analyze my Excel workbook and help me update its formulas and forecasts.",
"Compare revenue and costs across periods and create a report with charts.",
"Turn my data and findings into an editable PowerPoint presentation."
],
"developerName": "Julius AI",
"displayName": "Julius: Excel & Data Analysis",
"longDescription": "Julius AI helps analyze Excel spreadsheets, CSV files, and other files in your personal Julius library. Use Python to clean and combine datasets, run statistical analyses, build forecasts and financial models, and generate charts, HTML dashboards, reports, and editable PowerPoint presentations.\n\nImport a ChatGPT file attachment or use files already saved in Julius. Inspect inputs, run Python, and save results back to your personal library.\n\nSome examples of how you can use Julius:\n\nExcel and spreadsheet editing: Use Python to update formulas, combine worksheets, clean data, and revise budgets or forecasts in saved workbooks.\nStatistics and data analysis: Run regressions and hypothesis tests, identify outliers, and visualize relationships in your data.\nBusiness reporting: Compare revenue, costs, and performance across periods using your files, then generate charts, HTML dashboards, and reports.\nMarketing analytics: Analyze campaign exports saved as spreadsheets or CSV files and produce comparison reports and charts.\nPresentations: Use Python to turn datasets and analytical findings into editable PowerPoint files.\n\nCreate and revise text, code, and self-contained HTML files. When you request public sharing, publish a saved file and receive its public URL, or withdraw that link with unpublish.\n\nConnect your Julius account to get started.",
"logo": "./assets/julius-head-bust.png",
"privacyPolicyURL": "https://julius.ai/privacy",
"shortDescription": "Analyze data and build reports",
"supportURL": "https://julius.ai/support",
"termsOfServiceURL": "https://julius.ai/tos",
"websiteURL": "https://julius.ai"
},
"apps": "./.app.json"
}
},
"name": "julius",
"version": "1.1.0",
"$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json"
}
SHA-256: a3ffaa2f3f6c2f6b329387a186a089c2b1a3c448009d67bf258392c176f8efd9