The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs users to install langchain-mcp-adapters via pip. This is an official package maintained by LangChain (a trusted organization). It also includes examples using npx to fetch standard MCP servers like @modelcontextprotocol/server-filesystem.
[COMMAND_EXECUTION]: The documentation describes the stdio transport mechanism, which allows the adapter to launch local MCP server processes using shell commands. This is the intended primary functionality for local tool integration.
[PROMPT_INJECTION]: The skill enables loading prompts and resources from external MCP servers using functions like load_mcp_prompt and get_mcp_resource. This creates a surface for indirect prompt injection where data from an external server could potentially influence agent behavior.
Ingestion points: Data enters the context via load_mcp_resources, get_mcp_resource, and load_mcp_prompt functions described in SKILL.md.
Boundary markers: No specific delimiters or "ignore instructions" warnings are included in the provided code examples.
Capability inventory: The adapter supports subprocess execution (via stdio transport) and network operations (via sse, websocket, and streamable_http transports).
Sanitization: No evidence of content sanitization or validation is present in the skill's instructional examples.

langchain-mcp-adapters