loaditout.ai
SkillsPacksTrendingLeaderboardAPI DocsBlogSubmitRequestsCompareAgentsXPrivacyDisclaimer
{}loaditout.ai
Skills & MCPPacksBlog

mcp-token-auditor

MCP Tool

Ismail-2001/mcp-token-auditor

A high-performance, multi-agent observability engine designed for the Model Context Protocol (MCP). It provides a non-blocking, transparent proxy layer that implements deterministic token attribution, real-time context-window alerting, and heuristic-driven static analysis to optimize LLM metadata overhead at scale.

Install

$ npx loaditout add Ismail-2001/mcp-token-auditor

Platform-specific configuration:

.claude/settings.json
{
  "mcpServers": {
    "mcp-token-auditor": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-token-auditor"
      ]
    }
  }
}

Add the config above to .claude/settings.json under the mcpServers key.

About

🔎 MCP Token Auditor: Real-Time Performance intelligence

[](https://opensource.org/licenses/MIT) [](https://www.python.org/downloads/) [](https://modelcontextprotocol.io)

MCP Token Auditor is a production-grade, multi-agent proxy intelligence layer designed for high-throughput observability of MCP (Model Context Protocol) tool consumption. It acts as an transparent observer between MCP clients and servers, providing deterministic token counting, real-time alerting, and static analysis for tool optimization.

---

🏗️ System Architecture
graph TD
    Client[MCP Client] <--> ProxyAgent[Agent 2: Proxy Interceptor]
    ProxyAgent <--> Server[MCP Server]
    
    ProxyAgent -- "Ingest (Async Task)" --> Orchestrator[Agent 1: Orchestrator]
    
    subgraph "Intelligence Plane"
        Orchestrator --> Audit[Agent 3: Token Audit]
        Audit --> Alert[Agent 4: Alert Monitor]
        Audit --> Advisor[Agent 5: Compression Advisor]
        
        Audit[(SQLite Audit Log)]
        Advisor -- "Strategy" --> SharedSchema[Shared Schema Insights]
    end
    
    subgraph "Observability Plane"
        Audit -- "Stream" --> Dashboard[Agent 6: Dashboard Broadcast]
        Dashboard -- "WebSockets" --> UI[React/Web Dashboard]
    end

---

🚀 Key Features
  • ⚡ Zero-Latency Proxying: Decoupled audit logic using non-blocking background tasks ensures <5ms overhead on the critical path.
  • 🔢 Deterministic Counting: Uses tiktoken (o200k_base) with MD5-keyed caching for sub-millisecond token counts.
  • 🚨 Advanced Alerting: Real-time enforcement of CALL_SPIKE, SERVER_DRIFT, and context window thresholds.
  • 📉 Compression Advisor: Static analysis heuristics (Cloudflare code-mode, redundancy, deduplication) to reduce context

Tags

agentic-aifastapillm-cost-optimizationmcpobservabilityproxy-serverpythontoken-counting

Reviews

Loading reviews...

Quality Signals

0
Installs
Last updated23 days ago
Security: AREADME

Safety

Risk Levelmedium
Data Access
read
Network Accessnone

Details

Sourcegithub-crawl
Last commit3/25/2026
View on GitHub→

Embed Badge

[![Loaditout](https://loaditout.ai/api/badge/Ismail-2001/mcp-token-auditor)](https://loaditout.ai/skills/Ismail-2001/mcp-token-auditor)