mcp for VS Code
dev.safeprompt/mcp
Detect prompt injection, jailbreaks, and code injection in untrusted text before it reaches an LLM.
client:VS Code
transport:stdio
runtime:npm
Install mcp in VS Code
.vscode/mcp.json
{
"inputs": [
{
"type": "promptString",
"id": "safeprompt-api-key",
"description": "SAFEPROMPT_API_KEY",
"password": true
},
{
"type": "promptString",
"id": "safeprompt-provider",
"description": "SAFEPROMPT_PROVIDER",
"password": true
},
{
"type": "promptString",
"id": "safeprompt-user-ip",
"description": "SAFEPROMPT_USER_IP",
"password": true
}
],
"servers": {
"mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@safeprompt.dev/mcp"
],
"env": {
"SAFEPROMPT_API_KEY": "${input:safeprompt-api-key}",
"SAFEPROMPT_PROVIDER": "${input:safeprompt-provider}",
"SAFEPROMPT_USER_IP": "${input:safeprompt-user-ip}"
}
}
}
}VS Code prompts for each input the first time the server starts, so no secret is written to the file. VS Code docs