phoenix-uploads
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a mechanism for ingesting and processing untrusted user data via file uploads, creating a potential attack surface.
- Ingestion points: External data enters the application context through the
consume_uploaded_entriesfunction documented inSKILL.md. - Capability inventory: The provided code snippets utilize
File.mkdir_p!andFile.cp!to perform file system write operations. - Sanitization: The skill mitigates common file upload risks by including a
safe_filename/1helper that usesEcto.UUID.generate()to ensure unique, non-executable filenames, effectively preventing path traversal attacks. - Boundary markers: The instructions do not define markers or validation rules for the internal content of the uploaded files.
Audit Metadata