The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill accesses authentication tokens for the TikHub service.
It attempts to read tokens from environment variables TIKHUB_WECHAT_MCP_TOKEN or TIKHUB_MCP_TOKEN.
It fallbacks to reading credentials from the ~/.codex/config.toml file.
API response headers and bodies, which may contain session IDs or tokens, are written to temporary files in /tmp during processing.
[COMMAND_EXECUTION]: The script executes external system commands.
It uses child_process.spawnSync to call curl for interacting with the MCP server.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to external infrastructure.
It communicates with https://mcp.tikhub.io/wechat/mcp.
It uses hardcoded Cloudflare IP addresses (104.26.2.227, 104.26.3.227, 172.67.75.115) for direct resolution during the connection process.
[PROMPT_INJECTION]: The skill processes untrusted external data, which could contain instructions intended to influence the agent's behavior.
Ingestion points: Search result metadata (titles, author names) retrieved from the WeChat Channels API.
Boundary markers: None. Results are printed directly to the output stream without encapsulation or protective instructions.
Capability inventory: The skill has the ability to execute shell commands and read specific configuration files on the host system.
Sanitization: It implements a stripTags function to filter HTML tags and entities from the ingested data before display.

wechat-channels-search