btst-integration
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the user or agent to execute CLI commands using
@btst/cli, specifically for generating database schemas (generate) and running migrations (migrate). These are standard operations for the library's workflow.\n- [EXTERNAL_DOWNLOADS]: The instructions involve installing multiple packages from the@btst/namespace (e.g.,@btst/stack,@btst/adapter-prisma) and third-party libraries such as@tanstack/react-queryand@ai-sdk/openai. These resources are consistent with the skill's purpose and the vendor's domain.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection as it guides the agent in generating application code and database migrations based on the user's existing project structure.\n - Ingestion points: The agent ingests information about the user's project files, such as
lib/stack.ts,app/api/data/[[...all]]/route.ts, and global stylesheets.\n - Boundary markers: None specific to the templates; the skill assumes a trusted developer environment.\n
- Capability inventory: The skill facilitates file creation (Next.js/React Router routes), shell command execution (
npx @btst/cli), and database schema modifications.\n - Sanitization: None; the agent is expected to correctly interpolate templates into the specific application context.
Audit Metadata