sshepherd
Pass
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates remote server management by executing a locally-installed 'sshepherd' binary. The agent is instructed to use an absolute path provided by the user to invoke various DevOps operations, such as service control and log tailing.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it can read remote logs, database tables, and configuration files. This external data could contain malicious prompts intended to manipulate the agent's behavior.
- Ingestion points: Operations like 'logs tail', 'db tables', and 'config read' in SKILL.md bring remote, untrusted content into the agent's context.
- Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings to help the agent distinguish between data and instructions.
- Capability inventory: The agent has the capability to execute system commands through the 'sshepherd' CLI tool.
- Sanitization: There is no mention of sanitizing or validating the content retrieved from remote servers before it is processed by the agent.
Audit Metadata