hydra-head-operator
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION] (SAFE): The skill leverages the
Bashtool to run specific binaries (hydra-node,hydra-tui,docker,curl) necessary for Hydra Head operations. These commands are limited to the intended administrative scope. - [DATA_EXFILTRATION] (SAFE): Although the skill interacts with sensitive key files (
cardano.sk,hydra.sk), it does not attempt to transmit data to external domains. Instructions include setting restrictive file permissions (chmod 600) and network usage is confined tolocalhost. - [PROMPT_INJECTION] (SAFE): No patterns for direct or indirect prompt injection were detected. The use of
disable-model-invocation: trueensures that all tasks are performed under direct human supervision, reducing the risk of autonomous exploitation.
Audit Metadata