remote-script-exec
Warn
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONPERSISTENCE
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill defines a protocol for dynamically generating scripts on a local machine and executing them on remote hosts via SSH.\n- [COMMAND_EXECUTION]: The instructions explicitly suggest using the
-o StrictHostKeyChecking=noflag for SSH connections to unfamiliar nodes. This disables standard host identity verification, making the connection vulnerable to man-in-the-middle (MITM) attacks.\n- [PRIVILEGE_ESCALATION]: The skill directs the agent to use/root/directories for persistent work on remote nodes, which encourages the use of administrative privileges.\n- [INDIRECT_PROMPT_INJECTION]: The skill involves processing data from remote nodes, creating a potential surface for indirect injection attacks.\n - Ingestion points: Remote artifacts, logs, and data files are fetched via
scpfor local analysis by the agent.\n - Boundary markers: No boundary markers or instructions to ignore embedded commands within the fetched data are present.\n
- Capability inventory: The agent possesses extensive shell execution and file management capabilities across local and remote environments.\n
- Sanitization: The skill does not mention any sanitization or validation of content retrieved from remote nodes before it is interpreted.\n- [DATA_EXFILTRATION]: Provides a standard "Fetch back" procedure for transferring files from remote nodes to the local environment.\n- [PERSISTENCE]: Describes methods for maintaining long-term presence on remote nodes, such as using
nohupfor background tasks and creating dedicated persistent directories.
Audit Metadata