The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill demonstrates how to configure tool servers that are downloaded and executed at runtime using npx. These examples point to official packages from the @modelcontextprotocol organization, which is the standard provider for these services.
[COMMAND_EXECUTION]: The documentation describes the stdio transport, which allows the agent to spawn subprocesses to run local MCP servers as part of its toolset.
[PROMPT_INJECTION]: The skill facilitates the ingestion of data from external sources (such as GitHub issues or Notion pages) via MCP tools. This creates a surface for indirect prompt injection where instructions hidden in the external data could attempt to influence the agent's behavior.
Ingestion points: External data retrieved from services like GitHub and Notion through configured MCP servers.
Boundary markers: Not specified in the configuration examples provided in the documentation.
Capability inventory: The setup enables the agent to execute shell commands via the stdio transport and perform network operations via http and sse transports.
Sanitization: The provided configuration snippets do not include explicit sanitization or filtering logic for tool outputs.

cursor-mcp