neon-auth
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- EXTERNAL_DOWNLOADS (LOW): The skill references multiple markdown files and rule sets from GitHub (neondatabase-labs). Although this organization is not on the predefined trusted list, the files serve as instructional context for the agent and are appropriate for a service-specific integration tool.
- COMMAND_EXECUTION (LOW): The skill utilizes the
bashtool to perform project configuration and package management tasks. This level of access is expected given the skill's primary function of modifying application code and dependencies. - DATA_EXPOSURE (SAFE): No sensitive credentials or hardcoded keys were found. The provided templates correctly leverage environment variables for authentication configuration.
- PROMPT_INJECTION (LOW): The skill presents an indirect prompt injection surface (Category 8) because it uses
read_fileto ingest local project data (e.g.,tsconfig.json) to detect framework context. Evidence: 1. Ingestion points: Local project configuration files. 2. Boundary markers: Absent. 3. Capability inventory:bash,write, andread_file. 4. Sanitization: No explicit validation of file content before processing.
Audit Metadata