design
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes context from previous conversation steps or user-provided arguments to generate technical specifications. While this is an ingestion point for untrusted data, the skill's capabilities are scoped to reading existing repository files and producing a design document, presenting minimal risk of automated exploitation.
- [EXTERNAL_DOWNLOADS]: The instructions describe how to design dependency management for the MegaLinter project using Renovate-compatible comments and Dockerfiles. These are architectural guidelines for the target project and do not involve the skill itself performing unauthorized downloads or remote code execution.
Audit Metadata