execute-probe
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external specifications that control agent actions. 1. Ingestion points: Defined in SKILL.md as probe_spec and target. 2. Boundary markers: No delimiters defined to isolate input data. 3. Capability inventory: The skill description enables runtime tool selection for executing attacks. 4. Sanitization: No input validation or filtering is specified.
- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute commands and tools. 1. Evidence: The procedure explicitly mandates running probes and attacks using dynamically selected tools.
Audit Metadata