chart-assistant

Pass

Audited by Gen Agent Trust Hub on Nov 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes JavaScript utility modules (scripts/insert-image.mjs and scripts/deliver-image.mjs) that invoke the lark-cli command-line tool. This tool is used to interact with Feishu's document and media APIs for uploading images and modifying document blocks. The scripts use child_process.execFile to execute commands with controlled arguments.
  • [EXTERNAL_DOWNLOADS]: The export-image.mjs script performs network operations to https://charteasy.visactor.com to render chart data into images. This domain is owned by the skill's author, and the activity is limited to the skill's core purpose of chart visualization.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external, potentially untrusted sources such as Feishu Sheets and Base databases, presenting an indirect prompt injection surface.
  • Ingestion points: Data is ingested through standard data values and external source URLs defined in the chart configurations and referenced in references/sources-and-sync.md.
  • Boundary markers: The instructions in SKILL.md specifically warn the agent to disregard any instructions found within the input materials or data sources and to focus strictly on chart generation.
  • Capability inventory: The skill can create and edit Feishu documents and transmit data to a remote rendering service for image conversion.
  • Sanitization: The skill provides a structural validation script (scripts/common-option-validation.mjs) to verify the schema of chart configurations and specific business protocols before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 3, 2026, 09:33 AM