power-word-stun
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions to execute SIGSTOP signals and account lockouts. These are inherently disruptive administrative actions.
- [INDIRECT_PROMPT_INJECTION]: The skill processes a user-defined target for disruption without providing clear boundary markers or sanitization steps.
- Ingestion points: Target process or account identifier in the command invocation.
- Boundary markers: No delimiters are specified to isolate the target input from the agent's instructions.
- Capability inventory: Execution of process control signals and administrative account holds.
- Sanitization: No input validation is defined for the target string.
Audit Metadata