# Agent Guardrail

Action-level policy enforcement for AI agents — control what agents DO, not just what they say. Evaluate actions against configurable security policies before execution.

## Quick Start

```bash
# Connect this server (installs CLI if needed)
npx -y @smithery/cli@latest mcp add eren-solutions/agent-guardrail

# Browse available tools
npx -y @smithery/cli@latest tool list eren-solutions/agent-guardrail

# Get full schema for a tool
npx -y @smithery/cli@latest tool get eren-solutions/agent-guardrail evaluate_action

# Call a tool
npx -y @smithery/cli@latest tool call eren-solutions/agent-guardrail evaluate_action '{}'
```

## Direct MCP Connection

Endpoint: `https://agent-guardrail--eren-solutions.run.tools`

## Tools (6)

- `evaluate_action` — Evaluate whether an agent action is allowed by the guardrail policy. Call this BEFORE executing any tool, shell command…
- `register_agent` — Register a new agent with the guardrail system. Must be called before evaluate_action.
- `list_agents` — List all agents currently registered in the guardrail system. Shows active and killed agents.
- `get_stats` — Retrieve aggregate statistics from the guardrail system including evaluation counts, policy data, and agent metrics.
- `kill_agent` — Emergency kill-switch: immediately block ALL actions for an agent. Use when an agent is misbehaving or compromised.
- `unkill_agent` — Re-enable a previously killed agent, restoring its ability to have actions evaluated.

```bash
# Get full input/output schema for a tool
npx -y @smithery/cli@latest tool get eren-solutions/agent-guardrail <tool-name>
```

## Prompts (2)

- `setup_guardrail` — Step-by-step guide to set up guardrail enforcement for a new agent.
- `security_best_practices` — Security best practices for AI agent governance.