roboflow-cloud-storage
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external cloud storage buckets (AWS S3, GCS) which represents a potential attack surface for indirect prompt injection.
- Ingestion points: External data enters the Roboflow environment via tools like
datasource_validate(which performslistFiles,headFile, andgetFileoperations) anddatasource_triggerwhich initiates mirror jobs. - Boundary markers: The documentation does not specify boundary markers or instructions to ignore embedded instructions within the processed bucket objects.
- Capability inventory: The skill possesses capabilities to manage cloud credentials, define datasource configurations, and trigger recurring mirror jobs.
- Sanitization: There is no mention of sanitization or filtering of object metadata or content during the mirroring process.
- [SAFE]: The skill implements a secure credential handling design. It explicitly states that sensitive secrets (API keys, service account JSON, passwords) are collected out-of-band via MCP elicitation or manual entry URLs, ensuring they never appear as tool arguments, in the model context, or within chat transcripts. Credentials returned by the API are masked server-side.
Audit Metadata