hexo-blog-manager

The skill fragment presents a coherent workflow for Hexo post management, image hosting, and AI-generated covers, with explicit steps for encoding assets, uploading via GitHub API, and CDN link usage. However, several concerns elevate risk: (1) reliance on a non-official CDN domain (jsdelivr.007666.xyz) which could indicate tampering or misdirection; (2) handling of large base64 payloads via CLI inputs, which can leak credentials or overwhelm logs if not properly secured; (3) external service dependence (GitHub, Hugging Face) without explicit credential handling visible in the fragment; (4) a user-approval gate that could be overridden or bypassed in automation; (5) potential data exposure of assets/prompts to third-party services depending on prompt content and metadata. Overall, the footprint is functionally aligned with the stated Hexo publishing objective but is moderately suspicious due to domain, external dependencies, and data flow patterns. Recommendation: treat as SUSPICIOUS (securityRisk ~0.55) and require explicit validation of CDN domain authenticity, strict handling of temporary payloads, and secure management of tokens/API keys. Potential mitigations include pinning CDN to a trusted domain, avoiding long-lived temp files, and providing explicit security hardening notes in the docs.