alterlab-gtex
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from the external GTEx Portal API, creating a potential surface for indirect prompt injection.
- Ingestion points: The script
scripts/query_gtex.pyand code examples inSKILL.mdfetch genomic data from various endpoints underhttps://gtexportal.org/api/v2/. - Boundary markers: The skill processes structured JSON data but does not employ explicit delimiters or specialized instructions to the agent to ignore any natural language instructions that might be embedded within data fields like gene symbols or tissue metadata.
- Capability inventory: The skill utilizes
Bash(python:*)andWebFetchtools to retrieve and process API results. - Sanitization: No explicit sanitization or filtering of the API string content is performed before the data is presented to the agent's context.
- [EXTERNAL_DOWNLOADS]: The skill documents and performs downloads from external domains to retrieve scientific datasets.
- Evidence:
SKILL.mdprovideswgetcommands for downloading large-scale genomic data fromstorage.googleapis.com/adult-gtex/. - Evidence:
scripts/query_gtex.pyandSKILL.mdcode snippets use therequestslibrary to queryhttps://gtexportal.org/api/v2/. - Context: The targets are well-known scientific resources and cloud infrastructure providers, and the downloads are directly related to the skill's primary function of genomic data analysis.
Audit Metadata