data-visualization
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation recommends installing external dependencies using
npx skills add inference-sh/skills@agent-toolsand lists several other related skills from the same source. - [COMMAND_EXECUTION]: The skill requires the
Bashtool to executeinfshCLI commands, includinginfsh loginfor authentication andinfsh app runfor task execution. - [REMOTE_CODE_EXECUTION]: The provided visualization recipes involve sending Python scripts and HTML snippets to remote applications (
infsh/python-executorandinfsh/html-to-image) for processing and rendering. - [PROMPT_INJECTION]: A potential surface for indirect prompt injection exists because the skill encourages generating Python code for visualizations based on data; if an agent interpolates untrusted external data into these code blocks without sanitization, it could lead to unintended code execution within the remote environment.
Audit Metadata