tooluniverse-noncoding-rna
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download large biological datasets from established academic repositories including TargetScan (targetscan.org) and miRTarBase (cuhk.edu.cn). These are standard resources in the bioinformatics community.
- [DYNAMIC_EXECUTION]: The skill includes Python scripts that use "requests" and "pandas" to programmatically download, unzip, and process data files. This allows the agent to handle data that is too large for direct API access or model context.
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the ingestion of external biological identifiers.
- Ingestion points: Biological identifiers (miRNA IDs, gene symbols) provided by users or retrieved from databases are used as inputs for subsequent tool calls.
- Boundary markers: None explicitly defined in the skill instructions.
- Capability inventory: The skill uses tools for literature search (PubMed), database queries (miRBase, RNAcentral), and local Python execution for data processing.
- Sanitization: The skill does not specify sanitization or validation of the biological identifiers before passing them to external tools.
Audit Metadata