The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill accepts natural language input from users to query a database, which could potentially contain malicious instructions intended to manipulate the agent.
Ingestion points: The script scripts/run_query.py takes user input through the query_words command-line argument.
Boundary markers: No explicit delimiters or instructions are used to separate the user query from the surrounding script logic or to warn the processing engine about untrusted data.
Capability inventory: The skill is capable of executing Python scripts (scripts/run_query.py) and performing read operations on a Neo4j knowledge graph containing historical reports and transaction data.
Sanitization: The user-supplied input is joined and passed directly to the internal query function without visible sanitization or filtering to prevent prompt injection or query manipulation.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute a local Python script (scripts/run_query.py) that handles natural language processing and database interaction.

graph-query