The Agent Skills Directory

DATA_EXFILTRATION (LOW): The scripts batch_gene_lookup.py, fetch_gene_data.py, and query_gene.py perform network requests to ncbi.nlm.nih.gov. While NCBI is a reputable government scientific resource, it is not included in the predefined whitelist of trusted domains for exfiltration analysis. No access to sensitive local files (e.g., SSH keys or credentials) was detected.
PROMPT_INJECTION (LOW): This skill is vulnerable to Indirect Prompt Injection (Category 8) because it ingests and processes data from external NCBI API responses and user-provided gene list files. 1. Ingestion points: scripts/batch_gene_lookup.py (lines 101, 142), scripts/fetch_gene_data.py (lines 79, 114, 144), and scripts/query_gene.py (lines 46, 79, 113). 2. Boundary markers: Absent; external data is presented to the agent without delimiters or instructions to ignore embedded commands. 3. Capability inventory: Network access via urllib.request and local file writing in scripts/batch_gene_lookup.py (line 258). 4. Sanitization: Standard JSON/XML parsing is performed, but no natural language sanitization or filtering is applied to the retrieved data before it is presented to the agent.

gene-database