technical-blog-writing

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

• Third-party content exposure detected (low risk: 0.10). SKILL.md documents using the inference.sh CLI with user-provided inputs (e.g., exa/search --input query strings and html-to-image --input HTML) to generate content, but it does not describe any runtime workflow that reads outsider-authored free text from external feeds/queues without explicitly supplied content.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes remote apps (e.g., belt app run infsh/python-executor) which send and execute user-provided code on the inference.sh platform, so the inference.sh endpoint (https://inference.sh) is a runtime external dependency that executes remote code.