slang
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from web search results and user-supplied manuscripts, creating a surface for potential injection attacks.
- Ingestion points: The agent retrieves data from external platforms such as TikTok, Douyin, Xiaohongshu, Weibo, and Bilibili via the WebSearch tool, and it also accepts user-provided text for 'body check' (体检) analysis.
- Boundary markers: The instructions do not define clear boundaries or provide explicit directives for the agent to ignore instructions that may be embedded within the external search data or the user manuscripts.
- Capability inventory: The skill has access to the WebSearch capability for external data retrieval and references a local guide file (references/slang-guide.md).
- Sanitization: There are no specified procedures for sanitizing, filtering, or validating the content retrieved from external sources before it is analyzed or incorporated into the agent's workflow.
Audit Metadata