process-watch
Pass
Audited by Gen Agent Trust Hub on Mar 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a kill command that enables the termination of system processes. This functionality uses the psutil library to send termination signals (SIGTERM or SIGKILL) to processes identified by the user or agent.
- Evidence: The kill function in scripts/process-watch.py implements this capability.
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it retrieves and processes process metadata, such as process names and command lines, which are controlled by external entities on the system.
- Ingestion points: The skill uses psutil.process_iter() in scripts/process-watch.py to fetch process information.
- Boundary markers: None; the retrieved strings are included in the agent's context without delimiters or instructions to ignore embedded commands.
- Capability inventory: The skill has process termination capabilities and access to system-wide diagnostic data.
- Sanitization: There is no evidence of sanitization or validation of process names or command-line strings before they are processed.
Audit Metadata