planetscale-readonly-inventory
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes the
pscaleCLI tool and makes API calls to interact with the database environment and retrieve metadata. - Evidence: usage of
pscale database show,pscale api <path>, andpscale branch connections topcommands. - [DATA_EXFILTRATION]: The skill is instructed to access and record sensitive infrastructure configuration files and role metadata.
- Evidence: Instructions to "Inspect repository files for frameworks, ORMs, migrations, SQL tagging, and connection config" and "Inspect Terraform or other infrastructure-as-code definitions for PlanetScale roles, backups, backup policies..."
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external, potentially untrusted sources which could be manipulated to influence agent behavior.
- Ingestion points: Repository files (connection configs, migration files, Terraform templates) and database telemetry (SQLCommenter tags, query patterns).
- Boundary markers: The instructions do not specify the use of delimiters or clear separation between untrusted data and agent instructions.
- Capability inventory: The skill utilizes the
pscaleCLI and PlanetScale API to retrieve database state and metadata. - Sanitization: There is no evidence of logic to sanitize or validate the content of repository files or database tags before processing.
Audit Metadata