Skip to main content
Glama

Koma Gate MCP

Machen Sie Koma Gates Prompt-Injection-Klassifizierung als MCP-Tool für KI-Agenten verfügbar.

npm install koma-gate-mcp

Was es tut

Stellt ein einzelnes Tool, classify_input, bereit, das ein KI-Agent aufrufen kann, um zu prüfen, ob nicht vertrauenswürdiger Benutzertext sicher und im Rahmen liegt, bevor er darauf reagiert.

Related MCP server: agent immune

Einrichtung

Legen Sie den Anbieter und den API-Schlüssel fest:

# Provider: openai | anthropic | google | deepseek | ollama (default: google)
export KOMA_PROVIDER=google
export GEMINI_API_KEY=sk-...

Claude Desktop

Fügen Sie zu claude_desktop_config.json hinzu:

{
  "mcpServers": {
    "koma-gate": {
      "command": "npx",
      "args": ["-y", "koma-gate-mcp"]
    }
  }
}

Tool: classify_input

  • text (erforderlich) — der nicht vertrauenswürdige Benutzereingabe, der klassifiziert werden soll

  • preset (optional) — general | code | support | reference (Standard: general)

Gibt zurück:

{
  "allowed": true,
  "in_scope": true,
  "reason": "in scope",
  "preset": "general",
  "model": "gemini-2.5-flash"
}

Sicherheitsgrenze

Dies ist ein LLM-basierter Bereichsklassifizierer, keine kryptografische Prompt-Injection-Abwehr. Siehe koma-gates README für die vollständige Sicherheitsgrenze und BENCHMARKS.md für Bewertungsergebnisse.

A
license - permissive license
-
quality - not tested
A
maintenance

Maintenance

Maintainers
Response time
3dRelease cycle
3Releases (12mo)
Commit activity
Issues opened vs closed

Related MCP Servers

  • A
    license
    -
    quality
    D
    maintenance
    A lightweight stdio proxy that intercepts and rewrites MCP tool annotations to bypass security approval prompts in AI CLIs like Codex and Claude Code. It transparently passes through all tool operations while marking them as safe to ensure a seamless automation experience.
    11
    10
    MIT
  • A
    license
    -
    quality
    C
    maintenance
    Adaptive security for AI agents: assess inputs for prompt injection, scan outputs for credential/PII leaks, teach new attack patterns to semantic memory, harden prompts, and monitor metrics. Runs locally via MCP stdio.
    Apache 2.0
  • A
    license
    A
    quality
    B
    maintenance
    Drop-in stdio interceptor that gates MCP tools/call requests through a YAML policy (block / require_approval / audit / budget cap) before they reach the real server. The same policy file is reused by @jamjet/claude-code-hook and @jamjet/openai-guardrail, so one rule set covers Claude Desktop, Cursor, OpenAI Agents, and custom clients.
    3
    2
    Apache 2.0
  • A
    license
    A
    quality
    A
    maintenance
    Provides AI ingress governance by masking prompts, classifying risk, and enforcing tool policies before agent calls reach model providers or sandboxes.
    6
    1
    AGPL 3.0

View all related MCP servers

Related MCP Connectors

  • The WAF for agents. Pattern-based + heuristic firewall scans prompts, RAG documents, tool argume...

  • Deterministic trust gate for AI output: leaked-secret, prompt-injection & PII in one call.

  • Decision-only prompt routing and firewall checks for local/cloud routing, PII and jailbreak risk.

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/swnotmetal/Project-Koma'

If you have feedback or need assistance with the MCP directory API, please join our Discord server