matlab-list-products
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from local
prodcontents.jsonandVersionInfo.xmlfiles which could potentially contain malicious instructions if the files on the host system are compromised. - Ingestion points: Reads content from
{matlabroot}/appdata/prodcontents.json,{supportpkgroot}/appdata/prodcontents.json, and{matlabroot}/VersionInfo.xml(SKILL.md). - Boundary markers: None present to distinguish between file data and agent instructions during processing.
- Capability inventory: Includes filesystem read operations to extract product and release information. No network or write capabilities are requested.
- Sanitization: The skill performs basic cleanup of product names by stripping version and platform tokens, but does not sanitize the content against embedded instructions.
Audit Metadata