primekg
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from a local CSV file (kg.csv) and passes the contents (node names, relationship types, and metadata) directly to the agent's context. This creates a surface where malicious instructions embedded in the knowledge graph data could influence the agent's behavior. \n
- Ingestion points: The kg.csv file is loaded in scripts/query_primekg.py and accessed by search_nodes, get_neighbors, and get_disease_context functions. \n
- Boundary markers: None. The data is converted to dictionaries and returned as JSON-like structures without explicit delimiting or instructions to ignore embedded commands. \n
- Capability inventory: The skill has read access to the local filesystem at a specific hardcoded path (/mnt/c/Users/eamon/Documents/Data/PrimeKG/kg.csv). It does not perform network operations or file writes. \n
- Sanitization: No sanitization or validation is performed on the strings retrieved from the CSV file before they are returned to the agent context.
Audit Metadata