The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities detected. The skill provides standard instructions for using the official langchain-neo4j library.
[EXTERNAL_DOWNLOADS]: Fetches the langchain-neo4j package from the official Python Package Index. The package is maintained by a trusted organization.
[COMMAND_EXECUTION]: Includes standard Docker commands to initialize a local Neo4j instance for development and testing.
[PROMPT_INJECTION]: The skill uses GraphCypherQAChain, which generates Cypher queries from natural language. This surface is noted in the 'Security Note' section, correctly advising the use of read-only credentials and highlighting the mandatory allow_dangerous_requests safety flag.
Ingestion points: User-provided natural language queries processed by GraphCypherQAChain.invoke in SKILL.md.
Boundary markers: None explicitly defined in the provided code snippets.
Capability inventory: The skill can perform database reads and writes via the Neo4jGraph instance, which is capable of executing arbitrary Cypher queries.
Sanitization: LangChain's internal logic and the mandatory allow_dangerous_requests=True flag serve as the primary security controls and user acknowledgement for this capability.

langchain-neo4j