auto-skill-creation
Warn
Audited by Gen Agent Trust Hub on Apr 1, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is vulnerable to path traversal. The 'name' argument is used directly in path construction without validation, which could allow writing files outside the intended skills directory.
- [COMMAND_EXECUTION]: The skill performs dynamic code generation by interpolating unsanitized strings into Python templates. This allows for code injection if malicious input is provided, as the generated files are intended to be executed or tested.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. Evidence: 1. Ingestion points: 'name' and 'description' arguments derived from agent intent. 2. Boundary markers: None. 3. Capability inventory: File system writing and dynamic module loading via importlib in generated tests. 4. Sanitization: None.
Audit Metadata