Odel
mcp server

mcp server

Local
@correctover1PythonUpdated 1mo ago

Real-time LLM output verification with 6-dim contract validation and auto-failover across providers.

Correctover - MCP Runtime Security & Agent Fault Diagnosis

PyPI npm License

Featured in awesome-mcp-servers Correctover MCP server

We are the runtime security layer for the MCP ecosystem. Featured in awesome-mcp-servers — the definitive MCP server directory.

Correctover is the first MCP server purpose-built for agent fault diagnosis & self-healing. It monitors tool calls across your LLM providers, detects failures in real-time, and auto-repairs them using a knowledge base of 215 fault patterns and 19 CVEs — all at 22μs P50 latency.

What It Does

CapabilityDescription
🔍 Fault DiagnosisInstantly diagnose why your agent's tool call failed — timeout, auth, schema, rate limit, provider down
🔧 Auto Self-Healing97.4% of detected faults are automatically repaired with zero human intervention
🔄 Multi-Provider FailoverSeamlessly failover across 9+ LLM providers when one goes down
📊 Fault Pattern MatchingMatch errors against 215 known fault patterns from 150K+ empirical data points
🔒 Runtime SecurityIntercept RCE, SSRF, and cloud credential hijacking attempts at the MCP layer
🔑 BYOKBring Your Own Key — works with Anthropic, OpenAI, Mistral, Google, Cohere and more

Pricing

Free forever for individual developers. Scale when you need to.

PlanPriceCalls/monthWhat's included
Free$01,500 (50/day)All 215 fault patterns, basic diagnosis
Starter$49/mo5,000Full diagnosis + repair suggestions
Pro$199/mo25,000+ Custom fault rules + batch API
Team$499/mo100,000+ Team dashboard + audit logs
EnterpriseCustomUnlimited+ Private deployment + custom fault DB + SLA

Pay-per-use: $0.05/call after free tier. No commit, cancel anytime.

Quick Start

Cursor / Claude Desktop / Windsurf

Add to your MCP config:

{
  "mcpServers": {
    "correctover": {
      "command": "npx",
      "args": ["-y", "correctover-mcp-server"],
      "env": {
        "ANTHROPIC_API_KEY": "your-key-here"
      }
    }
  }
}

MCP Tools

Once connected, you get 3 tools:

  • diagnose_error — Paste an error, get root cause + fix suggestion in seconds
  • get_fault_pattern — Look up a specific fault pattern by ID or keyword
  • get_repair_suggestion — Get step-by-step repair instructions for a detected fault

Why Correctover?

The Problem

MCP agents fail constantly. Tool call timeouts, auth errors, schema mismatches, provider outages, rate limits — the list goes on. Developers waste hours debugging errors that have known solutions.

Our Solution

We've built the largest empirical MCP fault database: 150,000+ real-world failure cases across 32 MCP frameworks, categorized into 215 distinct fault patterns with 97.4% automatic self-heal rate.

When your agent hits an error, Correctover:

  1. Detects the fault pattern in 22μs (P50)
  2. Classifies it against our 215-pattern knowledge base
  3. Repairs it automatically (or suggests a fix)
  4. Fails over to a healthy provider if needed

The Numbers

  • 215 fault patterns cataloged
  • 19 CVEs documented
  • 150K+ empirical data points
  • 97.4% automatic self-heal rate
  • 22μs P50 detection latency
  • 32 frameworks analyzed
  • 5,100+ monthly downloads (PyPI + npm)

CCS: The Open Standard

Correctover is the reference implementation of CCS (Runtime Security Conformance Standard for MCP) — a vendor-neutral open standard for MCP runtime security.

Framework Integration

We actively contribute security research to the MCP ecosystem:

FrameworkContributionStatus
CrewAISecurity Issue #6411, #6545Open
LangGraphSecurity Issue #8308Open
AgentScope (Alibaba)Security Issue #2069In Review
FastMCPCommunity PRMerged
AstrBotSecurity PR #8975Open

Security Research

Our fault database powers active security research:

  • GHSAs: 4 published (1 Accepted, 3 draft)
  • Fault Taxonomy: 215 patterns + 19 CVEs across 32 frameworks
  • Bug Bounty: Active submissions to huntr, HackerOne, MSRC

Links

License

Apache-2.0


Built for the MCP ecosystem. Securing agents, one fault at a time.