csvglow

Generate beautiful interactive HTML dashboards from CSV/Excel files with smart insights, auto-detected charts, correlations, and statistics.

Security Scan

The skill is internally consistent: it is an instruction-only wrapper that expects a local csvglow CLI to produce HTML dashboards; no unrelated credentials or installs are requested — but you should verify the csvglow binary you install and avoid using the server mode unless you trust it.

Purpose & Capability

Name/description match the declared requirement: the skill simply delegates to a 'csvglow' CLI to generate HTML dashboards from CSV/Excel files. The required binary is proportional to the stated purpose.

Instruction Scope

SKILL.md instructs the agent to run the csvglow CLI on user files and optionally open the generated HTML. That stays within the described purpose. One noteworthy instruction: csvglow has an '--mcp' (server) mode mentioned — running that would start a network service and could expose data beyond a local HTML output, so it expands the runtime scope and should be used with caution.

Install Mechanism

There is no install spec in the skill (instruction-only). That reduces surface risk in the skill itself, but it means the user/agent must obtain the csvglow binary externally. The SKILL.md/homepage points to a GitHub repo — verify releases and checksums before installing from third-party sources.

Credentials

No environment variables, credentials, or config paths are requested. This is proportionate for a local data-visualization CLI.

Persistence & Privilege

always is false and the skill does not request persistent privileges or modify other skills. Autonomous invocation is allowed (platform default) but not an additional privilege here.

Assessment

This skill is a thin wrapper that expects you to have the csvglow CLI installed. Before installing or running it: 1) Verify the csvglow binary's source (prefer official releases, package managers, or signed releases on the GitHub repo) to avoid installing malicious binaries. 2) Do not run the '--mcp' or any server mode unless you understand and trust it — server mode may expose data over the network. 3) When generating dashboards, the tool will read local CSV/Excel files (as intended); avoid feeding sensitive or confidential data unless you trust the binary and its provenance. 4) If you want stronger assurance, inspect the upstream repo code and release artifacts or run the CLI in a sandboxed environment first.

Generate a beautiful, interactive HTML dashboard from any CSV or Excel file. Auto-detects column types and relationships, generates smart multi-column insights, histograms, bar charts, cross-analysis crosstabs, time series, correlation heatmaps, scatter plots, and a sortable data table — all in a single self-contained HTML file with a dark gradient theme.

Use this skill when the user wants to:

• Visualize a CSV or Excel file
• Create a dashboard from tabular data
• Explore or analyze a dataset visually
• Generate charts from a spreadsheet
• Get insights about their data

Usage

csvglow <file>                       # Generate dashboard and open in browser
csvglow data.csv                     # CSV file
csvglow report.xlsx                  # Excel file
csvglow data.csv -o dashboard.html   # Custom output path
csvglow data.csv --no-open           # Don't auto-open browser

What it generates

• Smart insights: Multi-column narrative analysis that cross-references metrics (e.g. "Gadget Y has the highest discount yet lowest revenue despite lowest cost — consider discontinuing")
• Summary stats: Row count, column count, data types, missing values
• Numeric columns: Histograms with stats sidebar (mean, median, std, quartiles, outlier count)
• Categorical columns: Gradient bar charts of top values
• Cross analysis: Automatic categorical x numeric crosstabs with overall mean lines
• Date columns: Time series line charts with area fill
• Correlation heatmap: Auto-detected correlations between numeric columns
• Scatter plots: Auto-generated for highly correlated pairs (|r| > 0.7)
• Data table: Sortable, filterable preview of first 1000 rows
• Copy button: Each chart has a "Copy" button for pasting into slide decks

Notes

• Output is a self-contained HTML file — no server, no CDN, works offline
• Supports CSV, TSV, XLS, and XLSX files
• Auto-detects delimiters and encodings
• Large files (100k+ rows) are smart-sampled for visualizations while keeping full stats
• Also available as an MCP server: csvglow --mcp