feat: add TWZRD Agent Intel MCP agent trust verification example

[twzrd-sol]

Summary

Adds examples/mcp_agent_trust_twzrd.py — a working example demonstrating streamablehttp_client with a real-world remote MCP server.

The existing examples/mcp_tool_runner.py shows stdio-based MCP. This example shows the streamable-http transport pattern using TWZRD Agent Intel, a free MCP server that returns on-chain trust scores for autonomous Solana agents.

What the example demonstrates

• Connecting to a remote streamable-http MCP server (no local process)
• Using async_mcp_tool() + tool_runner() with a real public endpoint
• Practical use case: verifying agent trust before routing value in an x402 or multi-agent workflow

Usage

pip install "anthropic[mcp]"
python examples/mcp_agent_trust_twzrd.py

No API key required for the TWZRD tools. The score_agent, resolve_agent, and preflight_check tools are free at https://intel.twzrd.xyz/mcp.

Relationship to existing example

	mcp_tool_runner.py	mcp_agent_trust_twzrd.py (new)
Transport	stdio (npx process)	streamable-http (remote URL)
Server	filesystem (local)	TWZRD Agent Intel (always-on)
Requires local install	Yes (@modelcontextprotocol/server-filesystem)	No