The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a Python helper script (personal_wechat.py) to interface with a local Wisdom service API via Bash. This is the primary intended function of the skill for managing WeChat data.
[DATA_EXPOSURE]: The skill requires PERSONALWECHAT_API_TOKEN and PERSONALWECHAT_BASE_URL to be provided via environment variables. The instructions and code explicitly include warnings never to print, log, or echo the secret token, following security best practices for credential management.
[EXTERNAL_DOWNLOADS]: The skill references a self-hosted Wisdom service at a default IP (http://82.156.126.14:8000), but uses this only as a template for the user's own BYOC (Bring Your Own Connector) configuration.
[PROMPT_INJECTION]: The skill includes clear 'Safety Rules' and 'Gated' message sending instructions that explicitly forbid the AI from sending messages without clear, human-in-the-loop approval, mitigating potential autonomous injection risks.
[REMOTE_CODE_EXECUTION]: The skill includes an unattended.py module to verify if the execution environment (AceDataCloud scheduled tasks) has pre-authorized the skill for unattended writes, preventing unauthorized automated actions.

personal-wechat