database-scout
Pass
Audited by Gen Agent Trust Hub on Nov 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/db_explorer.pyexecutes SQL queries provided via command-line arguments. It mitigates the risk of unauthorized database modification by using the?mode=roflag for SQLite connections and setting the PostgreSQL session toREAD ONLY. Furthermore, it validates SQL input using regex to allow onlySELECT,WITH,EXPLAIN,PRAGMA, andSHOWstatements while blocking a comprehensive list of dangerous keywords. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external databases, which is a potential vector for indirect prompt injection if the database content contains malicious instructions.
- Ingestion points: Data is ingested via table names, column schemas, and row previews fetched from SQLite files or PostgreSQL instances in
scripts/db_explorer.py. - Boundary markers: There are no specific delimiters used in the console output to separate data from instructions, which is standard for CLI-based tools.
- Capability inventory: The script performs file system reads (for SQLite) and network communication (for PostgreSQL), executing SQL through standard library interfaces (
sqlite3,psycopg2). - Sanitization: The tool uses robust identifier escaping via
_quote_idand thepsycopg2.sqlmodule, alongside its SQL keyword filtering, to maintain security during execution.
Audit Metadata