The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill clones its primary codebase and supporting assets from a GitHub repository (github.com/oaker-io/wewrite.git). This is a standard installation procedure for this tool suite.
[COMMAND_EXECUTION]: The skill executes a series of Python scripts located in the scripts/ and toolkit/ directories via subprocess.run to perform various pipeline tasks, such as fetching hotspots, generating images, and publishing drafts.
[CREDENTIALS_UNSAFE]: The skill requires several sensitive credentials (WeChat AppID/Secret and AI provider API keys). It follows safe practices by instructing users to utilize environment variables or a separate configuration file rather than hardcoding secrets directly into the skill code.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from external trending topics (Weibo, Baidu, Toutiao) via scripts/fetch_hotspots.py which is then processed by the agent to create articles.
Ingestion points: scripts/fetch_hotspots.py (fetches real-time trending topics from Weibo, Toutiao, and Baidu).
Boundary markers: No explicit instruction delimiters or boundary markers are defined in the SKILL.md for the scraped content.
Capability inventory: subprocess.run (script execution), toolkit/publisher.py (WeChat API interaction), and file system writes to output/.
Sanitization: No explicit sanitization or validation of the scraped hotspot data is described before it is provided to the agent context.

wewrite-wechat-ai-publishing