mastra-file-agents
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a legitimate code refactoring process that translates Mastra agent definitions into the recommended file-based directory structure. It correctly handles various configuration types like static and dynamic instructions, memory instances, and custom tools.
- [COMMAND_EXECUTION]: The instructions include running
npx mastra devandnpx mastra build. These are official CLI tools for the Mastra framework and are used appropriately for verification within the developer's environment. - [PROMPT_INJECTION]: The skill processes existing source code to facilitate migration, which constitutes an indirect prompt injection surface. Ingestion points: Reads agent definitions from
src/mastra/index.tsand related project files. Boundary markers: Relies on structural code parsing and framework-specific patterns. Capability inventory: Creates new directories and writes TypeScript and Markdown files to the local filesystem. Sanitization: Uses predefined mapping logic for code translation rather than arbitrary string interpolation. Assessment: The behavior is inherent to its purpose as a refactoring tool and operates safely within the project scope.
Audit Metadata