dropbox
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes external data from a user's Dropbox account, which can be controlled by third parties.\n
- Ingestion points: The agent can search, read, and download files using
dropbox call-toolas described inSKILL.md, introducing potentially malicious content into the agent's context.\n - Boundary markers: The instructions do not define delimiters (like XML tags or specific markdown blocks) or provide clear directives for the agent to treat content from Dropbox as untrusted data.\n
- Capability inventory: The agent has powerful capabilities, including creating, modifying, and deleting files (
content.delete), and managing sharing settings (sharing.manage), which could be abused if an injection is successful.\n - Sanitization: There are no instructions for sanitizing, filtering, or escaping the retrieved file content before the agent processes it.
Audit Metadata