ai-hive-simulate-li-shanglong
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content such as links and transcripts, which presents a surface for indirect prompt injection.
- Ingestion points: External URLs and text inputs provided for content style analysis.
- Boundary markers: The skill instructions recommend a structured approach with specific disclaimers to separate external data from AI reasoning.
- Capability inventory: Included scripts (scripts/ai_hive_creator.py and scripts/ai_hive_mcp.py) can perform network requests and write local files.
- Sanitization: Relies on prompt-level instructions rather than programmatic input filtering.
- [COMMAND_EXECUTION]: The skill includes Python scripts intended for execution to manage content plans and interact with the AI-HIVE MCP service.
- [EXTERNAL_DOWNLOADS]: The provided scripts perform network operations to the vendor's domain at ai-hive.iclip.cn to retrieve metadata and interface with the generation platform.
Audit Metadata