single-cell-downstream-analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly requires downloading and ingesting public third-party assets that the agent will use to drive analysis (e.g., "git clone https://github.com/CSB5/CaDRReS-Sc", ov.utils.download_CaDRReS_model(), ov.utils.download_GDSC_data(), ov.utils.get_gene_annotation() pointing to GENCODE GTFs, and SCENIC cisTarget database downloads listed in SKILL.md), which are open web sources whose content is consumed and can materially influence tool execution and decisions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs cloning and using the CaDRReS-Sc GitHub repository (git clone https://github.com/CSB5/CaDRReS-Sc) as a prerequisite and passes scriptpath='CaDRReS-Sc' to ov.single.Drug_Response, meaning remote code from that URL is fetched at runtime and executed as a required dependency.