encode-database
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes structured and unstructured data from the ENCODE Portal and SCREEN APIs, which could serve as a vector for indirect prompt injection if the external data sources were compromised.
- Ingestion points: API responses retrieved via requests.get and requests.post from encodeproject.org and api.encodeproject.org in SKILL.md.
- Boundary markers: No specific delimiting or 'ignore instruction' markers are applied to the external genomics data.
- Capability inventory: The skill has network access (requests), file write capabilities (urllib.request.urlretrieve, pd.to_csv), and data decompression (gzip).
- Sanitization: The skill does not implement specific sanitization of the biological metadata before display or storage.
- [EXTERNAL_DOWNLOADS]: The skill facilitates the download of biological data files (BED, bigWig) from the ENCODE Project's official domain (encodeproject.org). These are retrieved from a well-known scientific service and are intended for research analysis.
Audit Metadata