Trentina MCP
A secure MCP gateway and quarantine for AI agent traffic that provides three-layer prompt injection defense. It stops prompt injection at every ingress, shrinks context window size through compression and tool filtering, and enforces policy via literal request and response guards. It supports OAuth/DCR authentication and per-agent profiles for architectural flexibility.
Package and access
Local package with source-attributed installation metadata. Koskamo does not execute packages during review.
What you can do with it
- Defend against prompt injection attacks at the gateway layer
- Reduce token usage by compressing responses and filtering tools
- Enforce semantic request and response guards via English rules
- Manage per-agent profiles with distinct backends and credentials
Source and maintenance
- Maintenance checked
Packages and installation
Source-provided package data. Koskamo has not executed or live-probed this package.
mcp-trentina-crunchtools 0.5.0
{
"version": "0.5.0",
"transport": "stdio",
"identifier": "mcp-trentina-crunchtools",
"registryType": "pypi"
}Tools
Reported by the attributed source; not independently observed by Koskamo.
Connect from your client
Each snippet configures the same source-provided package; replace placeholders before use.
Cursor
{
"mcpServers": {
"trentina-mcp": {
"command": "uvx",
"args": [
"mcp-trentina-crunchtools==0.5.0"
]
}
}
}Cline
{
"mcpServers": {
"trentina-mcp": {
"command": "uvx",
"args": [
"mcp-trentina-crunchtools==0.5.0"
]
}
}
}VS Code
{
"servers": {
"trentina-mcp": {
"type": "stdio",
"command": "uvx",
"args": [
"mcp-trentina-crunchtools==0.5.0"
]
}
}
}Claude Desktop
{
"mcpServers": {
"trentina-mcp": {
"command": "uvx",
"args": [
"mcp-trentina-crunchtools==0.5.0"
]
}
}
}Let your agent find its tools
Find servers like Trentina MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.