The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data and external schema files to generate and execute Cypher queries, creating a potential attack surface that is thoroughly mitigated by design.
Ingestion points: Processes external schema definitions in <db-name>-schema.json and user-controlled data via LOAD CSV commands.
Boundary markers: Implements a strict 'Write execution gate' requiring EXPLAIN validation and explicit human confirmation before any write operation is performed.
Capability inventory: Capable of generating and facilitating the execution of Cypher queries, including reading, writing, and database administrative procedures.
Sanitization: Mandates the use of $parameters for all queries and explicitly warns against unsafe string interpolation for untrusted values in dynamic execution contexts.
[COMMAND_EXECUTION]: The skill provides utility scripts and CLI examples intended for database interaction.
Evidence: Includes scripts/generate_schema.py and scripts/define_schema.py which allow the agent or user to fetch or define database schemas. It also provides a curl example for the Neo4j HTTP API using placeholders for credentials.
Note: These utilities are standard administrative tools provided by the author to support the skill's primary function and do not represent a security risk.

neo4j-cypher-skill