broadcast-event
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill's event summarization instructions create an indirect prompt injection surface. It suggests using a prompt template to summarize tool inputs and contents without using boundary markers or sanitization. Malicious instructions embedded in the data processed by the agent (such as tool outputs) could potentially manipulate the summarization behavior.\n
- Ingestion points: Tool inputs and content previews are processed in Step 2 of the instructions.\n
- Boundary markers: Absent in the summarization prompt template designed in Step 2.\n
- Capability inventory: The skill utilizes Read and Write tools which are part of its allowed-tools for implementing the design.\n
- Sanitization: No validation or escaping of interpolated variables is specified in the design pattern.\n- [COMMAND_EXECUTION]: The skill provides code templates for a Python WebSocket server and a JavaScript client, intended to be implemented using the Write tool. While provided as design patterns, this involves the generation of executable scripts on the local system.
Audit Metadata