image-upload
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests local image files which are then uploaded to external services, presenting a theoretical attack surface for indirect prompt injection.
- Ingestion points: Files are read from the local filesystem using
fs.readFilein thereaderfunction withinsrc/upload.ts. - Boundary markers: The instructions do not define specific delimiters or warnings to prevent the agent from interpreting text found in filenames or within the image data (e.g., metadata or OCR) as commands.
- Capability inventory: The skill has the capability to perform network POST requests to external hosting providers (such as catbox.moe, imgbb.com, and imgur.com) via
src/utils/fetch.ts. - Sanitization: The skill implements basic sanitization by verifying the existence of the file, checking the file size against provider limits, and performing MIME type detection using magic bytes to ensure only supported image formats are processed.
Audit Metadata