The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to clone its entire codebase from a third-party GitHub repository (https://github.com/oaker-io/wewrite.git) which is not associated with the skill author or a trusted organization.
[REMOTE_CODE_EXECUTION]: After cloning the external repository, the skill executes the downloaded Python scripts (scripts/*.py) and toolkit components (toolkit/*.py) to perform its primary functions.
[COMMAND_EXECUTION]: The skill uses extensive shell command execution to run various pipeline steps, including fetching trending topics, SEO analysis, image generation, and publishing articles.
[INDIRECT_PROMPT_INJECTION]: The skill fetches real-time "hotspot" data and search suggestions from third-party platforms like Weibo, Baidu, and Toutiao. This untrusted data is fed into the AI to influence topic selection and article generation, creating an injection surface.
Ingestion points: scripts/fetch_hotspots.py (crawls Weibo, Toutiao, Baidu) and scripts/seo_keywords.py (crawls search suggestions).
Boundary markers: None identified; instructions do not tell the AI to ignore embedded commands in the fetched data.
Capability inventory: The skill has significant capabilities including file writing, network API access for image generation (OpenAI/Doubao), and full access to the WeChat Official Account API for publishing content.
Sanitization: No sanitization or validation of the externally fetched hotspot data is mentioned before it is processed by the AI.

wewrite-wechat-ai-publishing