html-erd-explorer
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided database schemas and table structures.
- Ingestion points: Processes database schema descriptions and data models provided by the user in the chat context.
- Boundary markers: No specific text delimiters are defined, but output is constrained to a structured HTML format.
- Capability inventory: Generates and writes .html files to the local filesystem and creates interactive JavaScript for browser-based exploration.
- Sanitization: The instructions explicitly mandate the use of
textContentandcreateElementinstead ofinnerHTMLto prevent Cross-Site Scripting (XSS) when rendering user-supplied table or column names. - [EXTERNAL_DOWNLOADS]: The skill references Google Fonts via standard
<link>tags, which is a well-known service used for styling. It also references a local shared assetsubmit-handler.jsfor platform-specific integration. - [COMMAND_EXECUTION]: The skill instructions include writing generated artifacts to the local disk, which is the intended core functionality for creating portable documentation.
Audit Metadata