The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the zsxq-cli binary to perform various topic management actions, such as searching, posting, and replying on the Knowledge Planet platform.
[PROMPT_INJECTION]: The skill contains an indirect prompt injection vulnerability surface.
Ingestion points: Untrusted platform data is retrieved through topic +search, topic +detail, and the get_topic_comments API call across all documentation files.
Boundary markers: No explicit delimiters or instructions are provided to the agent to help it distinguish between platform-retrieved content and its own operating instructions.
Capability inventory: The skill allows for significant write operations, including +create for new posts, +reply for comments, and +answer for answering questions, which could be misused if the agent obeys instructions found within external data.
Sanitization: No sanitization, validation, or filtering mechanisms for the platform-provided data are described in the skill's logic.

zsxq-topic