linux-admin
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed for system administration and utilizes shell commands, systemd, and firewall management tools (iptables, firewalld). The use of sudo and administrative commands such as iptables -F or iptables -P INPUT DROP are consistent with its primary purpose of Linux server management.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it is instructed to read and process potentially untrusted data such as system logs, network traffic captures, and external command outputs. 1. Ingestion points: journalctl (references/systemd-units.md), tcpdump, lsof, and dig (references/networking.md). 2. Boundary markers: The instructions do not explicitly mandate the use of delimiters when interpreting log or tool output. 3. Capability inventory: The skill has high privileges, including sudo access, the ability to write to system files via tee and redirection, and the power to modify network and firewall configurations. 4. Sanitization: There are no explicit instructions for the agent to sanitize or escape data read from logs before processing them. Given the administrative context, this risk is considered safe as it is inherent to the functionality of a sysadmin tool.
- [SAFE]: The static analysis hits are verified as false positives. The reverse_shell_pattern in references/networking.md is a legitimate port connectivity check using /dev/tcp for environments where nc might be absent. The dangerous_process_control finding in references/iptables-guide.md refers to standard firewall management commands (flushing chains and setting default policies). The skill also promotes security best practices, such as disabling root login and using secure script headers (set -euo pipefail).
Audit Metadata