component-designer
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill lacks input sanitization for user-provided data such as component names and purposes, creating a surface for indirect injection or path traversal.\n
- Ingestion points: User input for component details and names gathered in Step 2 of the workflow (SKILL.md).\n
- Boundary markers: None present in the markdown templates used for file creation.\n
- Capability inventory: Use of the
Writetool to create new files and theEdittool to modify existing project architecture files.\n - Sanitization: No input validation or escaping is observed; user strings are interpolated directly into file paths and document content.
Audit Metadata