gget
Warn
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION] (MEDIUM): High potential for Indirect Prompt Injection via untrusted biological data inputs.
- Ingestion points: The script
batch_sequence_analysis.py(line 15) reads FASTA files containing arbitrary sequence IDs and data. The scriptenrichment_pipeline.py(line 15) reads gene lists from CSV or text files. - Boundary markers: Absent. External data is parsed and passed directly into analysis functions without delimiters or instructions to ignore embedded commands.
- Capability inventory: The skill calls the
ggetlibrary which performs network operations to external genomic databases (BLAST, Ensembl, Enrichr). It also performs extensive file writing (to_csv,open('w')) and directory creation (mkdir). - Sanitization: Absent. No validation or escaping is performed on sequences, gene names, or identifiers before they are used in API calls or as part of generated filenames.
Audit Metadata