safeagent

About safeagent

Use safeagent with your agent

claude mcp add safeagent -- uvx safeagent-exec-guard

{
  "mcpServers": {
    "safeagent": {
      "command": "uvx",
      "args": [
        "safeagent-exec-guard"
      ]
    }
  }
}

Frequently asked questions

What is safeagent?

safeagent is an MCP (Model Context Protocol) server by azender1 in the Autonomous Agents category. Exactly-once execution guard for AI agents. Prevents duplicate payments and webhooks on retry.

How do I connect safeagent to Claude, Cursor, or another MCP client?

Install safeagent from its pypi package (safeagent-exec-guard) and register it under "mcpServers" in your client's MCP configuration — for example claude_desktop_config.json or Cursor's mcp.json — then restart the client.

Is safeagent open source?

Yes — the source code is public at github.com/azender1/safeagent, with 6 GitHub stars.

Can AI agents pay to use safeagent?

Not yet through Loomal — safeagent is listed as a free directory entry. If its maintainer verifies ownership, they can set per-call USDC pricing that agents pay over x402, with settlement on Base.