neo4j-kafka-skill
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFECOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references official download paths for the Neo4j Kafka Connector on GitHub and the Neo4j website. These are legitimate vendor resources required for installation and do not pose a security risk.
- [COMMAND_EXECUTION]: The README.md and SKILL.md files include standard installation commands using
curlandconfluent-hub. These are intended for the initial setup of the connector in a controlled environment. - [INDIRECT_PROMPT_INJECTION]: The skill manages the ingestion of Kafka messages into a Neo4j database, which is an attack surface for data poisoning. The skill mitigates this by using Cypher parameters to isolate data from query logic. * Ingestion points: Kafka topics specified in the sink connector configuration. * Boundary markers: The use of the
$eventsparameter and theUNWINDclause to bind external payloads to internal variables. * Capability inventory: Database write operations via Cypher and environment setup via Bash. * Sanitization: Relies on the Neo4j driver's internal parameterized query mechanism. - [CREDENTIALS_UNSAFE]: The documentation includes fields for Neo4j and Kafka authentication. However, it explicitly warns users against hardcoding credentials and provides examples of using secure Kafka Connect secret providers (e.g.,
${file:...}) for production use.
Audit Metadata