web-frameworks
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions to run local Python utility scripts (
nextjs-init.py,turborepo-migrate.py) and standard shell commands for project scaffolding and task management. These operations are restricted to project setup and development workflows. - [EXTERNAL_DOWNLOADS]: The skill fetches assets and dependencies from established, well-known services. This includes icon fonts from the RemixIcon CDN, analytics and mapping SDKs from Google and Meta, and various development tools via the NPM registry.
- [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface where the agent processes external data.
- Ingestion points: External API data is fetched within Server Components, such as in
app/posts/[slug]/page.tsx. - Boundary markers: Absent in the provided architectural templates.
- Capability inventory: Limited to UI rendering; the skill does not contain capabilities for arbitrary file-system writes, shell command execution, or network exfiltration of the fetched data.
- Sanitization: Relies on React's default text interpolation which automatically escapes string content, mitigating direct execution of untrusted content.
Audit Metadata