skill-opt-lite
Pass
Audited by Gen Agent Trust Hub on Aug 16, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill explicitly instructs the agent to execute arbitrary shell commands defined in the
test_commandfield of the inputtasksJSON file during the rollout and scoring phases to verify success. - [INDIRECT_PROMPT_INJECTION]: The skill processes external task data that contains both natural language instructions and executable code, which are used to drive agent behavior and modify skill files.
- Ingestion points: The
tasksJSON array provided as a primary input parameter in the rollout and validation stages. - Boundary markers: The skill does not define specific delimiters or 'ignore' instructions when processing the external task data.
- Capability inventory: The agent environment allows for shell command execution (
test_command) and file system operations (editing skill documents and writing to the.skillopt/directory). - Sanitization: No sanitization, filtering, or validation of the task instructions or shell commands is performed before they are used or executed.
Audit Metadata