cachyos-linux-assistant
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use standard Linux administrative tools such as
pacman,systemctl,journalctl, andip. It explicitly references the use ofsudofor package management and system configuration tasks, which is consistent with the skill's primary purpose of Linux administration. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection due to its core functionality.
- Ingestion points: The skill instructions direct the agent to read and analyze untrusted system data, specifically system logs (
journalctl), kernel logs (dmesg), and package manager outputs (pacman -Q). - Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded prompts within the data being processed.
- Capability inventory: The skill provides the agent with the ability to execute shell commands, manage system services, and install or remove software packages.
- Sanitization: No specific sanitization, filtering, or validation mechanisms for external log content are defined in the instructions.
Audit Metadata