runpod
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references Docker images hosted on GitHub Container Registry (a well-known service) under the account
conalmullan. These images are external to the skill author and represent a third-party dependency. - [DATA_EXFILTRATION]: Documentation indicates that the associated tools fall back to "free upload services" for file transfers when Cloudflare R2 is not configured. This creates a risk of sensitive user data (images, video, audio) being transmitted to untrusted or unspecified third-party platforms.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted media data for processing via cloud AI models, presenting an injection surface.
- Ingestion points: Media files (images, videos, audio) processed by scripts in the
tools/directory (e.g.,image_edit.py,upscale.py). - Boundary markers: The instructions do not define specific delimiters or instructions to ignore potential adversarial content embedded in processed media or model outputs.
- Capability inventory: The skill executes local scripts via
uv run, performs network operations to RunPod APIs, and handles file transfers including uploads to external services and downloads of results. - Sanitization: The skill does not mention explicit sanitization, validation, or filtering of the content extracted from or generated by the external AI models.
Audit Metadata