mxl-compile
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a compiler for JSON data that may be derived from untrusted sources, such as OCR data from user-provided images of forms.
- Ingestion points: The skill processes JSON content loaded from the file path provided in the
JsonPathargument within bothscripts/mxl-compile.ps1andscripts/mxl-compile.py. - Boundary markers: The skill does not utilize specific prompt delimiters, but the conversion logic strictly treats input as data for a structured XML format.
- Capability inventory: The skill has the capability to write XML files to arbitrary local paths defined by the
OutputPathparameter. - Sanitization: The scripts implement robust sanitization through dedicated XML escaping functions (
Esc-Xml,Esc-XmlText, andesc_xml_text). These functions are consistently applied to all cell text, parameters, and template strings before they are written to the output file, neutralizing potential XML injection attacks.
Audit Metadata