nosql-databases
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides code patterns for ingesting and processing external data from databases, which presents a surface for indirect prompt injection attacks.\n
- Ingestion points: Functions in SKILL.md such as insert_event and put_event process data dictionaries from external sources.\n
- Boundary markers: The provided snippets do not utilize delimiters or explicit instructions to separate data from instructions.\n
- Capability inventory: The skill facilitates database read and write access for MongoDB, Redis, DynamoDB, and Cassandra.\n
- Sanitization: Explicit data sanitization or content validation is not demonstrated in the implementation examples.\n- [SAFE]: The skill implementation follows multiple security best practices, including the use of prepared statements in Cassandra queries, connection pooling for Redis, and the use of yaml.safe_load in the validation script scripts/validate.py. Connection strings use local endpoints and do not contain hardcoded credentials or secrets.
Audit Metadata