neo4j-spark-skill
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the construction of Cypher queries through options like
queryandscript.N, which can ingest untrusted data from Spark DataFrames or external inputs, potentially leading to injection attacks if the data is not sanitized. - Ingestion points: The
queryoption in both read and write modes (e.g., accessing DataFrame row values viaevent.column_namein write queries) and thescript.Npre-write execution feature. - Boundary markers: The documentation does not explicitly define or enforce boundary markers, delimiters, or mandatory parameterization for user-provided data within Cypher templates.
- Capability inventory: The skill enables full CRUD operations on a Neo4j database, including the execution of arbitrary Cypher scripts before write operations, which could be exploited if injection occurs.
- Sanitization: There is no mention of specific sanitization, validation, or escaping techniques for external content before it is interpolated into query templates.
Audit Metadata