The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references the @apideck/mcp package and the vendor's official GitHub repository. It instructs users to use npx for execution, which is the standard distribution method for this type of service.
[COMMAND_EXECUTION]: The documentation includes instructions to run the command npx -y @apideck/mcp. This command facilitates the local execution of the vendor's MCP server code.
[DYNAMIC_EXECUTION]: The skill documents an experimental 'Code mode' feature that uses vm.createContext. It proactively warns developers to avoid using this in production environments due to potential security limitations of the sandbox implementation.
[INDIRECT_PROMPT_INJECTION]: By acting as a gateway to over 200 SaaS connectors, the skill manages a significant attack surface for indirect prompt injection from external data sources.
Ingestion points: Data flow from various SaaS platforms (Accounting, CRM, HRIS, etc.) via the Unified API tools.
Boundary markers: No specific delimiters or boundary markers are defined within the skill's instruction set.
Capability inventory: The server exposes 330 tools across multiple domains, involving file operations, network API calls, and data persistence.
Sanitization: The skill relies on the internal security architecture of the Apideck service and the downstream agent runtime for input sanitization.

apideck-mcp