sql-database-assistant
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest natural language descriptions from users to generate SQL queries and migration scripts, which creates a surface for indirect prompt injection attacks.
- Ingestion points: Untrusted user input is processed via the
changeparameter inscripts/migration_generator.pyand thequeryparameter inscripts/query_optimizer.py. - Boundary markers: There are no explicit delimiters or instructions provided in the skill metadata or scripts to help the agent distinguish between its core instructions and potentially malicious commands embedded in the user data.
- Capability inventory: The script
scripts/migration_generator.pycontains a file-writing capability through the--outputcommand-line argument, which allows the generated output to be saved to a local path. An attacker could potentially use prompt injection to influence the content and the target destination of this file write. - Sanitization: The provided scripts use regular expressions for parsing but do not perform validation on target file paths or sanitize user-supplied descriptions against adversarial instructions.
Audit Metadata