Koma Gate MCP

Classifies prompt injection, jailbreaks, and out-of-scope user input before agents act on it.

Connector by swnotmetal · GitHub account verified by the MCP Registry

IndexedMITLocal (stdio)★ 13

Install

Works with Claude Code, Codex, VS Code, Cursor, Windsurf, Claude Desktop. Read the repository before you give any tool access to your data.

npx agentmarket install connector/swnotmetal-koma-gate-mcp

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

About

Classifies prompt injection, jailbreaks, and out-of-scope user input before agents act on it.

Reviews

Write a review

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this server can do and what it asks for, from its tool list, install config and code. Checked Oct 26, 2026.

Tools: 4 read · 0 write · 0 deleteRead-only
Secrets: 4 (API keys, tokens or passwords)
Install scripts: None found
Runs as: Node.js package

Tools by what they do 4

Read from the source code; the running server may list different tools.

Read 4 · Look things up without changing anything.

batch_classify
classify_input
get_stats
list_presets

Secrets and settings it asks for

Secrets: values you paste in that let it act as you.

ANTHROPIC_API_KEY
DEEPSEEK_API_KEY
GEMINI_API_KEY
OPENAI_API_KEY

Other settings (environment variables or headers).

KOMA_MODEL
KOMA_PROVIDER

Network hosts it talks to 2

api.deepseek.com · found in code
github.com · found in code

Tools (4)

Read from the source code; the running server may differ. Descriptions are the server’s own words; hidden characters are shown as [U+…].

batch_classify
Classify multiple untrusted inputs in a single call. Returns per-input safety decisions. Useful for bulk-screening a queue of user messages before processing them.
classify_input
Check whether user input is safe and in-scope before an AI system processes it. Detects prompt injection, jailbreak attempts, instruction overrides, and off-topic requests. Use this to validate untrusted user text before passing it to an LLM or executing a tool.
get_stats
Return runtime statistics for the active guard instance: total classifications, allowed, rejected, cached, and error counts.
list_presets
List the available guard presets (domains) and their default models. Use this to discover which classification domain best matches your use case.

Compare and alternatives

Featured in

Signals

Stars: 13
Forks: 2
Open issues: 1
Last commit (this folder): Aug 18, 2026 (2 months ago)
Last release: Aug 15, 2026 (2 months ago)

From GitHub (fetched Oct 24, 2026).

Details

Registry name: io.github.swnotmetal/koma-gate-mcp
Version: 0.1.1
Auth: API key
Packages: npm koma-gate-mcp
License: MIT
Repository: github.com/swnotmetal/Project-Koma
Website: github.com/swnotmetal/Project-Koma/blob/main/packages/koma-gate-mcp/README.md
First seen: Oct 24, 2026
Last checked: Oct 24, 2026
Source: Official MCP Registry github-graphql