prisma
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the
@membranehq/clipackage globally via NPM. This is a standard installation of a tool provided by the skill's vendor to facilitate interaction with their platform. - [COMMAND_EXECUTION]: The skill utilizes the
membranecommand-line utility for essential operations, including user authentication, creating database connections, and running Prisma actions. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external Prisma databases, which creates a potential surface for indirect prompt injection if the processed records contain instructions that could influence the AI agent's behavior.
- Ingestion points: Data returned from the Prisma API via
membrane action runandmembrane requestcommands. - Boundary markers: No specific delimiters or safety instructions are defined for the data returned from Prisma.
- Capability inventory: The skill possesses the ability to execute shell commands through the Membrane CLI and perform network requests via the proxy.
- Sanitization: There is no evidence of data sanitization or validation performed on the records fetched from the external database.
Audit Metadata