jinko-trial
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data which creates a vulnerability surface for indirect prompt injection.\n
- Ingestion points: External data enters the agent context through trial results (TimeSeries and Scalar data) downloaded via the SDK and data tables attached to trials during setup.\n
- Boundary markers: The instructions do not define clear delimiters or include warnings to ignore instructions potentially embedded within the ingested data tables or trial results.\n
- Capability inventory: The skill utilizes the Jinkō SDK for network operations (creating trials, polling, and downloading data) and uses matplotlib to save generated plot figures to the filesystem.\n
- Sanitization: Content within the downloaded dataframes is not explicitly sanitized or validated for malicious instructions before being used by the agent for analysis or plotting.\n- [EXTERNAL_DOWNLOADS]: The skill directs the agent or user to install additional required tools from the vendor's GitHub repository (novainsilico/jinko-skills) if the jinko-sdk-setup skill is missing.
Audit Metadata