planetscale-vitess-safety-review
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external data from database telemetry, schema recommendations, and anomaly events, which creates a potential surface for indirect prompt injection if those sources are compromised.
- Ingestion points: External PlanetScale database metadata, query insights, and anomaly payloads (SKILL.md).
- Boundary markers: Absent; the instructions do not define specific delimiters for external data processing.
- Capability inventory: The skill is restricted to generating reports and proposals. It includes strict prohibitions against automated execution: 'Do not apply any changes', 'Never reshard... automatically', and 'Do not create branches without approval'.
- Sanitization: No specific sanitization or escaping of external database content is mentioned.
- [SAFE]: The skill is authored by a known vendor and focuses on operational best practices. It includes strong security guidance, such as advising against agents self-approving deploy requests and ensuring a human-in-the-loop for all database modifications.
Audit Metadata