The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts publish.js and publish_with_video.js use child_process.execFileSync to invoke the external @wenyan-md/cli tool (wenyan) to transform and upload Markdown content to WeChat.
[CREDENTIALS_UNSAFE]: The publishing scripts are designed to load sensitive API credentials (WECHAT_APP_ID, WECHAT_APP_SECRET) from environment variables or by parsing specific local files in the user's home directory, such as ~/.openclaw/workspace/TOOLS.md.
[DATA_EXFILTRATION]: The skill communicates with official WeChat API endpoints (api.weixin.qq.com) and Sogou search services (weixin.sogou.com) to search, download, and publish content. No unauthorized network destinations were identified.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it ingests untrusted data from the web which is then processed by the agent.
Ingestion points: Untrusted article content and search results are ingested via search_wechat.js and download.js.
Boundary markers: The skill does not implement explicit boundary markers or instructions for the agent to ignore embedded commands within the ingested content.
Capability inventory: The skill has the capability to write to the local file system and execute shell commands through the publisher scripts.
Sanitization: There is no evidence of sanitization or filtering of the web-sourced content before it is presented to the agent for rewriting or 'washing'.

wechat-toolkit