planetscale-postgres-safety-review
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed exclusively for auditing and recommending best practices. It contains numerous restrictive instructions (e.g., 'Do not apply changes', 'Do not create branches', 'Do not rotate roles') that prevent unauthorized modifications to the target infrastructure.
- [DATA_EXPOSURE]: The skill includes security-positive recommendations, such as using user-defined roles with least privilege and keeping database passwords out of Terraform state files.
- [INDIRECT_PROMPT_INJECTION]: The skill reviews external data such as SQL comments, query tags, and webhook metadata.
- Ingestion points: SQL comments (SQLCommenter tags), query patterns, and database metrics retrieved via the PlanetScale API (SKILL.md).
- Boundary markers: Not explicitly defined for the ingested data.
- Capability inventory: The skill's capabilities are limited to gathering information and generating recommendations. It explicitly forbids performing any 'write' or 'delete' operations (creating branches, migrations, role management, network changes, or restores) without human approval.
- Sanitization: Not explicitly documented for the reviewed query content, but the risk is mitigated by the lack of execution capabilities.
Audit Metadata