Skills
skills/automateyournetwork/netclaw/pyats-linux-system/Gen Agent Trust Hub

pyats-linux-system

Pass

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Provides functionality to execute arbitrary shell commands on remote Linux hosts via the pyats_run_linux_command tool, which is a powerful administrative primitive.
  • [PROMPT_INJECTION]: The skill relies on natural language instructions to restrict actions to read-only operations. These guardrails are not technically enforced and are susceptible to prompt injection or behavioral overrides.
  • [DATA_EXFILTRATION]: Facilitates inspection of sensitive system paths such as /etc, /var/log, and /home for diagnostic purposes, creating a potential surface for sensitive data exposure.
  • [PROMPT_INJECTION]: Vulnerable to indirect prompt injection due to the lack of technical boundaries around command execution. Ingestion points: command parameter in pyats_run_linux_command (SKILL.md). Boundary markers: Missing technical delimiters. Capability inventory: Remote command execution and filesystem access (SKILL.md). Sanitization: None programmatically implemented.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 18, 2026, 06:13 AM