bluebubbles

Pass

Audited by Gen Agent Trust Hub on Jul 16, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The sendAttachment action enables the agent to access files on the local filesystem via the path parameter and send them to external recipients. This capability could be abused to exfiltrate sensitive files if the agent is manipulated into accessing unauthorized paths.
  • [PROMPT_INJECTION]: The skill processes message content and interaction data from external sources without using boundary markers or sanitization, creating a surface for indirect prompt injection.
  • Ingestion points: Message bodies, recipient handles, and attachment file paths defined in SKILL.md.
  • Boundary markers: None present to distinguish between system instructions and untrusted external message content.
  • Capability inventory: File system access and network transmission capabilities integrated via the messaging channel.
  • Sanitization: No instructions provided for the validation or filtering of incoming data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 16, 2026, 09:13 PM
Security Audit — agent-trust-hub — bluebubbles