postgresql-best-practices
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFE
Full Analysis
- Sanitized Data Ingestion: The skill implements defensive practices against indirect prompt injection by explicitly instructing the agent to treat all database content as untrusted data rather than instructions. It also requires the agent to present target impacts and request user confirmation before executing state-changing CLI commands or modification-based SQL operations.
- Trusted External References: The skill refers to official certificate downloads from DigiCert, a well-known certificate authority, for the configuration of secure SSL/TLS connections. These references are documented neutrally and follow industry-standard security practices.
- Managed Service Guardrails: The instructions include a dedicated section on Managed Service Guardrails that prevents the agent from attempting unsupported operations on cloud-managed PostgreSQL instances (such as direct OS-level access or local file path modification), ensuring the agent operates within the constraints of the managed platform.
- Clear Execution Policy: The skill defines a strict shell execution policy for cloud CLI tools, requiring discovery before action and subscription verification, which helps prevent accidental operations on unintended cloud resources.
Audit Metadata