blink-backend
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses specific MCP tools (
blink_backend_status,blink_backend_deploy,blink_backend_logs) to interact with the Blink CLI for project management and deployment. These are authorized tools within the skill's defined context. - [EXTERNAL_DOWNLOADS]: The code examples reference the
@blinkdotnew/sdkandhonopackages. These are standard dependencies for the Blink platform and web development, respectively. - [CREDENTIALS_SAFE]: The instructions explicitly emphasize using auto-injected environment variables (
BLINK_SECRET_KEY) and project settings for third-party keys, warning the user never to hardcode secrets. - [INDIRECT_PROMPT_INJECTION]: The skill provides patterns for handling incoming webhooks and API requests which process untrusted external data. This represents the skill's primary intended use case for server-side processing.
Audit Metadata