baidu-milan-winter-olympics-2026
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The scripts (
milan-olympics.js,milan-news.js,milan-china-medals.js, andmilan-schedule.js) fetch data fromtiyu.baidu.com. This is a well-known service providing public sports information and is utilized here for its stated purpose. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external web pages, which is then processed and returned to the agent's context, creating a potential attack surface.
- Ingestion points: Remote content is fetched from
https://tiyu.baidu.com/in all four scripts. - Boundary markers: The script outputs do not include explicit delimiters or instructions to the agent to disregard potential commands embedded within the fetched data.
- Capability inventory: The skill uses the Node.js
httpsmodule to perform network requests but does not possess other high-privilege capabilities like file writing or arbitrary code execution. - Sanitization: The scripts implement basic sanitization by extracting specific fields using regular expressions and JSON parsing, which limits the raw content passed to the agent to a predefined schema.
Audit Metadata