aqua-communication
Warn
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [REMOTE_CODE_EXECUTION]: The installation process in
SKILL.mdandREADME.mduses a remote script execution pattern. It downloads an installer from the author's GitHub repository and executes it via bash. - [COMMAND_EXECUTION]: The installation instructions require the use of
sudoto run the installer, granting the script administrative privileges on the host system. - [PROMPT_INJECTION]: The skill introduces a surface for indirect prompt injection (Category 8) because its primary purpose is to receive and process messages from external peers.
- Ingestion points:
aqua inbox listandaqua inbox watchcommands inSKILL.mdwhich provide message content to the agent. - Boundary markers: Absent. Message content is not wrapped in isolation markers to prevent the agent from following instructions contained within the messages.
- Capability inventory: The agent uses the
aquaCLI to send and receive data across a peer-to-peer network. - Sanitization: Absent. The tool does not perform filtering or sanitization of incoming message payloads.
- [EXTERNAL_DOWNLOADS]: The skill fetches its core utility and installation scripts from
github.com/quailyquaily/aqua.
Audit Metadata