stream-react
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows secure development practices, specifically by prohibiting the agent from reading or editing sensitive environment files and delegating credential handling to the official Stream CLI. It also ensures that generated projects include a proper .gitignore file for security.
- [COMMAND_EXECUTION]: Shell commands are employed for project setup and development workflows, including scaffolding with npx shadcn and running the Next.js development server. These operations use specific allowed patterns and trusted tools.
- [EXTERNAL_DOWNLOADS]: The skill downloads project templates and the Stream SDKs from official registries and GitHub repositories managed by established organizations such as Vercel and Microsoft.
- [REMOTE_CODE_EXECUTION]: The skill utilizes npx to execute build tools like tsc and next. It also uses node -e for utility scripts to check for installed icon packages and update package configuration during scaffolding.
- [PROMPT_INJECTION]: The skill builds applications processing untrusted data via Chat (MessageList) and Feeds (useFeedActivities) ingestion points. While the skill lacks explicit LLM-facing boundary markers for code analysis, it provides sanitization patterns for user IDs and server-side authentication. The agent's capabilities include shell execution (npm, npx, node) and file read/write access.
Audit Metadata