The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests raw, unstructured meeting data and transcriptions provided by users, which creates a surface for indirect prompt injection. If the provided text contains malicious instructions, the agent might be manipulated during the formatting or document creation process.
Ingestion points: Raw meeting information, user-provided unstructured notes, and audio transcriptions gathered in Step 1 of the workflow.
Boundary markers: None identified. The instructions do not specify using delimiters or warnings to ignore instructions embedded within the meeting content.
Capability inventory: The skill uses yuque_create_doc from the yuque-mcp server to write content to a persistent external storage service.
Sanitization: The instructions recommend cleaning up filler words in transcriptions but do not include security-related sanitization or escaping of the meeting content before document creation.
[SAFE]: The skill uses yuque-mcp tools for their intended purpose of listing personal repositories and creating documents. All network activity and tool interactions are directed to the official Yuque platform, which is consistent with the skill's stated purpose of personal knowledge base management.

personal-meeting-notes