app-builder
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
Bashtool to perform project initialization and dependency management (e.g.,npx create-next-app,npm install). While standard for a scaffolding tool, this involves executing commands that incorporate project-specific parameters. - [EXTERNAL_DOWNLOADS]: The skill's templates instruct the agent to download and install various third-party packages and frameworks from official registries like NPM and PyPI. These are well-known technology sources.
- [PROMPT_INJECTION]: As an orchestrator, the skill ingests natural language user input to plan and generate code, creating a surface for indirect prompt injection or command injection.
- Ingestion points: User project descriptions enter the context through
SKILL.mdandproject-detection.md. - Boundary markers: The skill instructions do not specify any delimiters or safety warnings to delineate untrusted user input from the agent's core logic.
- Capability inventory: The skill has access to powerful tools including
Bash,Write, andEditto carry out generated plans. - Sanitization: There is no documented logic for sanitizing user-provided strings (such as project names or feature descriptions) before they are used in terminal commands or project templates.
Audit Metadata