technical-blog-writing

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill's Quick Start explicitly instructs running the infsh search app ("infsh app run exa/search --input '{ "query": "building REST API Node.js best practices 2024 tutorial" }'"), which clearly performs web-style research by fetching open/public third-party content that the agent is expected to read and use to inform writing decisions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly requires and invokes the inference.sh CLI (https://inference.sh) at runtime (e.g., "infsh app run infsh/python-executor" and other infsh app runs), which executes arbitrary remote code and/or uses remote apps to control generation, so this external URL is a required runtime dependency that executes code.