postgres
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external PostgreSQL databases, which serves as a potential vector for indirect prompt injection. If a database contains malicious content, it could attempt to manipulate the agent's behavior during the processing of query results.
- Ingestion points: SQL query results returned by the
scripts/query.pyscript to the agent. - Boundary markers: The documentation does not define specific delimiters or isolation instructions for the data returned from the database.
- Capability inventory: The skill has the ability to execute local scripts and establish network connections to PostgreSQL hosts.
- Sanitization: While the skill validates SQL syntax (restricting to SELECT, SHOW, EXPLAIN, etc.) and uses read-only sessions, it does not implement sanitization for the actual content of the data fetched from the database to prevent prompt injection.
Audit Metadata