typed-service-contracts
Pass
Audited by Gen Agent Trust Hub on Apr 22, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities were detected. The skill acts as a documentation and template resource for architectural best practices.
- [PROMPT_INJECTION]: The skill defines a pattern for handling external inputs which are subsequently used in file system operations. While this creates an attack surface for indirect prompt injection, the provided templates include robust sanitization.
- Ingestion points: Input schemas defined in spec.ts (e.g., MyTaskInputSchema).
- Boundary markers: Use of Zod for data parsing and validation.
- Capability inventory: Local file system access via the fs module in handler.ts (e.g., fs.existsSync).
- Sanitization: Implementation of SafePathSchema using Zod refinements to explicitly reject path traversal (..) sequences.
Audit Metadata