The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the uxc command-line utility for managing Notion MCP interactions. It defines a workflow for creating command aliases (uxc link) to streamline tool invocation.\n- [EXTERNAL_DOWNLOADS]: Network operations are directed towards mcp.notion.com/mcp, which is the official endpoint for Notion's Model Context Protocol service. This is a well-known service and the interaction is necessary for the skill's functionality.\n- [PROMPT_INJECTION]: The skill has an inherent surface for indirect prompt injection as it retrieves data from external Notion workspaces. This is addressed through instructions for strict JSON parsing and mandatory human-in-the-loop confirmation for all write-access operations. Evidence Chain: 1. Ingestion points: notion-fetch and notion-search in SKILL.md. 2. Boundary markers: Instruction to use JSON envelopes and parse stable fields. 3. Capability inventory: uxc link, uxc auth, and MCP execution. 4. Sanitization: Mandatory user confirmation for write actions and read-first workflows.

notion-mcp-skill