Add Armorer Guard

[armorer-labs]

Adds Armorer Guard to the tools and code section using the repository contribution format. It is a local Rust scanner that can inspect MCP and AI-agent tool-call arguments/results for prompt injection, credential leakage, exfiltration, and risky actions before they reach the model or execute.

[armorer-labs]

Quick context update for this MCP-security list: the latest Armorer Guard release added a first-class MCP proxy mode:

armorer-guard mcp-proxy -- npx your-mcp-server

It gates stdio MCP tools/call arguments before forwarding them to the wrapped server, with local checks for credential leakage, dangerous tool-call intent, exfiltration, prompt injection, and local policy matches. It returns structured JSON-RPC errors and makes no scanner network calls.

Demo: https://huggingface.co/spaces/armorer-labs/armorer-guard-demo
MCP proxy docs: https://github.com/ArmorerLabs/Armorer-Guard/blob/main/examples/mcp_proxy.md

Happy to change the category/description if there is a preferred taxonomy for runtime/proxy guardrails.