agent-guard-mcp

MCP server that detects and prevents infinite agent loops — the #1 reliability problem in agentic systems.

Provides circuit breakers, pattern detection, stuck-agent analysis, and recovery recommendations via the Model Context Protocol.

Install

npx agent-guard-mcp

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "agent-guard": {
      "command": "npx",
      "args": ["agent-guard-mcp"]
    }
  }
}

From source

git clone https://github.com/mdfifty50-boop/agent-guard-mcp.git
cd agent-guard-mcp
npm install
node src/index.js

Tools

register_agent

Register an agent for monitoring.

log_action

Log an agent action for loop detection. Maintains a rolling window of the last 50 actions.

Returns early warnings when repeated patterns are detected.

detect_loop

Check if an agent is stuck in an infinite loop.

Returns:

• is_stuck — boolean
• confidence — 0.0 to 1.0
• pattern — healthy, exact_repeat, low_diversity, or exact_repeat_and_low_diversity
• suggestion — human-readable recovery advice
• repeated_actions — list of repeated signatures with counts

set_circuit_breaker

Configure automatic intervention when action patterns repeat.

check_circuit_breaker

Pre-flight check before executing a tool. Call this BEFORE the actual tool call.

Returns proceed: false when the circuit breaker fires (block mode only).

get_stuck_report

Detailed analysis of why an agent is stuck.

Returns: action history, pattern analysis, token waste estimate, diversity ratio, top repeated patterns, and recovery recommendations.

get_health_dashboard

Overview of all monitored agents. No parameters.

Returns all agents sorted by risk score with their status (STUCK/HEALTHY), action counts, and circuit breaker config.

Resources

Usage Pattern

1. register_agent — at agent startup
2. Before each tool call:
   a. check_circuit_breaker — should this action proceed?
   b. Execute the tool
   c. log_action — record what happened
3. Periodically:
   - detect_loop — am I stuck?
   - get_health_dashboard — how are all agents doing?
4. On stuck detection:
   - get_stuck_report — what went wrong and how to recover

Tests

npm test

License

MIT