mcpdirectory.dev
9,129 servers · updated 23 Sept

Agent Sec MCP server

community Active last commit 53 d 
{
  "mcpServers": {
    "agentsec": {
      "type": "http",
      "url": "https://agentsec.kernora.ai/mcp"
    }
  }
}
file .mcp.json (in your project root)version 0.1.0

What it does

Agent Sec exposes a read-only security baseline of 15 rules covering secrets, injection, supply-chain, and other risks, cited to OWASP, CWE, and regulations. It allows AI coding agents to check proposed actions against known-good practices and self-correct. The service runs as a stateless HTTP endpoint with no installation or signup required.

  • Check whether a proposed command or action violates security rules before the agent executes it, receiving citations to the relevant OWASP or CWE standard.
  • Retrieve the full curated security baseline at the start of a session so the agent can reason about compliance as it generates code.
  • Self-host the baseline on your own infrastructure for air-gapped or privacy-critical environments, then point agents to that URL instead.

Tools this server exposes

ToolDoes
get_security_baselineRetrieve the full set of 15 security rules with their citations and descriptions.
check_actionPass an action or command string and receive back which baseline rules apply to it, with citations.

Before you enable it

An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.