chart-image

Pass

Audited by Gen Agent Trust Hub on Nov 22, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill leverages reputable and well-known libraries, specifically Vega-Lite and Sharp, for data visualization and image processing. Its code is clear, well-documented, and focused on its stated purpose.
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts untrusted user data in JSON and CSV formats to generate charts. This creates a potential surface where malicious instructions could be embedded in data labels or titles, intended for an agent that visualizes the output. This risk is inherent to the functionality of data visualization tools.
  • Ingestion points: The scripts/chart.mjs file ingests data via the --data, --csv, and --csv-file command-line arguments, as well as through standard input (stdin).
  • Boundary markers: None are present in the provided scripts to delimit user data from instructions.
  • Capability inventory: The skill has the capability to read and write files using fs.readFileSync and fs.writeFileSync within the scripts/chart.mjs utility.
  • Sanitization: The skill uses JSON.parse and an internal CSV parser to process input data, which provides basic structural validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 22, 2026, 01:11 AM