z-go-llm-streaming
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill provides technical guidelines and code examples for Go developers. The code implements standard patterns for LLM integration, such as interface decoration, retry logic, and SSE streaming. All operations are local or directed at well-known LLM provider interfaces.\n- [INDIRECT_PROMPT_INJECTION]: The skill describes how to process external data from LLM providers and users.\n
- Ingestion points:
RequestandMessagestructs inSKILL.mddefine the data surface.\n - Boundary markers: Not explicitly defined in these architectural templates.\n
- Capability inventory: The skill utilizes network capabilities to perform its primary function of LLM communication.\n
- Sanitization: Not applicable to the provided transport-layer code snippets.\n- [COMMAND_EXECUTION]: The verification section includes
curlcommands intended for manual local testing by the developer. These commands targetlocalhostand do not involve remote script execution or malicious activity.
Audit Metadata