planetscale-autonomous-execution-mode
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a mechanism for 'standing authorization' where the agent reads instructions from a local file to bypass interactive approvals.
- Ingestion points: The agent is instructed to read authorization artifacts from
AGENTS.mdor a dedicated local authorization file at the start of every run in scheduled or webhook-triggered contexts. - Capability inventory: The skill facilitates significant database mutations, including schema changes (ADD INDEX, ADD COLUMN), branch management, and deployment requests (Class B, C, and D operations).
- Boundary markers: There are no instructions for verifying the integrity, signature, or source of the
AGENTS.mdor authorization files before the agent acts on them. - Sanitization: The skill does not specify any sanitization or validation logic to ensure that the authorization artifact itself has not been tampered with to include malicious intents.
Audit Metadata