prism for VS Code
io.github.tulipe1735/prism
Complete web tasks in the user's own Chrome via an observe-decide-act browser agent.
client:VS Code
transport:stdio
runtime:npm
Install prism in VS Code
.vscode/mcp.json
{
"inputs": [
{
"type": "promptString",
"id": "typesafe-api-key",
"description": "TYPESAFE_API_KEY",
"password": true
},
{
"type": "promptString",
"id": "text-model-api-key",
"description": "TEXT_MODEL_API_KEY",
"password": true
},
{
"type": "promptString",
"id": "text-model-base-url",
"description": "TEXT_MODEL_BASE_URL",
"password": true
},
{
"type": "promptString",
"id": "text-model",
"description": "TEXT_MODEL",
"password": true
}
],
"servers": {
"prism": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@tulipe1735/prism"
],
"env": {
"TYPESAFE_API_KEY": "${input:typesafe-api-key}",
"TEXT_MODEL_API_KEY": "${input:text-model-api-key}",
"TEXT_MODEL_BASE_URL": "${input:text-model-base-url}",
"TEXT_MODEL": "${input:text-model}"
}
}
}
}VS Code prompts for each input the first time the server starts, so no secret is written to the file. VS Code docs