The Agent Skills Directory

[COMMAND_EXECUTION]: The skill interacts with the ZSXQ platform using the zsxq-cli tool to perform topic management tasks such as searching, viewing details, and posting content. This is the intended and documented use of the CLI tool.\n- [PROMPT_INJECTION]: The skill uses directive language such as 'CRITICAL' and 'MUST' to ensure the agent reads necessary authentication and error handling rules in a shared configuration file. These instructions are functional for operation and do not attempt to bypass agent safety guidelines.\n- [DATA_EXFILTRATION]: While the skill retrieves content from an external platform, it does so for the purpose of displaying it to the user. There are no patterns suggesting the exfiltration of sensitive data to unauthorized third-party domains.\n- [SAFE]: The skill incorporates strong security mitigation by instructing the agent to confirm all public write operations (posting, replying, answering) with the user before execution, ensuring human-in-the-loop control for irreversible actions.

zsxq-topic