syncfusion-java-word
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents features for ingesting and processing external data formats such as HTML, XML, and JSON, which creates a surface for indirect prompt injection attacks.\n
- Ingestion points: Reference files document techniques to ingest data from external sources, specifically
references/html-conversions.md(HTML ingestion),references/xml-conversions.md(XML ingestion), andreferences/mail-merge.md(JSON and XML data sources).\n - Boundary markers: The documentation includes validation methods like
isValidXHTML()inreferences/html-conversions.mdto check schema, but lacks explicit instructions for the AI to treat ingested data as untrusted or to ignore embedded instructions.\n - Capability inventory: The documented Java code snippets involve extensive file system operations including
WordDocument.open()andWordDocument.save(), and Java I/O stream handling across multiple reference files, representing a functional capability that could be exploited if an injection occurs.\n - Sanitization: The library provides
XHTMLValidationTypefor schema validation inreferences/html-conversions.md, helping to maintain data structure integrity.
Audit Metadata