Topic · Model Context Protocol

MCP servers

A Model Context Protocol server exposes tools that an AI client can call. I run a public, read-only endpoint here for my profile, project descriptions, tools, and TechTide AI service links. It does not execute project code or return private records.

Connect to the live endpoint

The endpoint is public and requires no authentication. Use the configuration below with a compatible MCP client. You can inspect the declared tools at /.mcp/list-tools; compatibility depends on the client.

MCP endpoint

{
  "mcpServers": {
    "alexcinovoj": {
      "url": "https://alexcinovoj.dev/mcp"
    }
  }
}

Tools exposed

get_profile returns my published bio, role, and location.
list_projects returns project descriptions and available repository links.
list_mcp_servers returns MCP server examples, repository links, and setup commands.
list_stack returns the tools I use, grouped by purpose.
list_engagements returns TechTide AI service names and links, without prices.

What the endpoint can and cannot do

I declare the available tools explicitly. The endpoint does not pass requests to a code execution environment.
The tools return read-only information and do not modify project data.
The handlers return fixed public content. They do not read private records or credentials.
The annotations describe intended tool behavior; handlers must enforce access. The linked example explains permission checks for other MCP servers.

MCP permission scoping pattern Agent governance

Related projects

ClawKeeper

ClawKeeper checks proposed Claude agent actions against explicit rules before a model call. Autonomous spend defaults to zero, and the database enforces an append-only audit log.

TechTide Harness Kit

TechTide Harness Kit uses JSON Schema to declare skill permissions, risk tiers, network access, and approval requirements. Automated checks validate declarations and run prompt-injection fixtures.