Running a restricted sandboxed agent with access to host MCP
What is the best practice to run a restricted agent (very limited tools, one mcp tool or one specific exec command and not full exec).
Does using elevated inherits the exec limitations set on the agent?
{
"agents": {
"summarizer": {
"security": "allowlist",
"allowlist": [
{ "pattern": "**/mcporter" }
]
}
}
}
Will this in combination with elevated allow the sub agent to call the mcp only on the host ?
Does using elevated inherits the exec limitations set on the agent?
{
"agents": {
"summarizer": {
"security": "allowlist",
"allowlist": [
{ "pattern": "**/mcporter" }
]
}
}
}
Will this in combination with elevated allow the sub agent to call the mcp only on the host ?
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
83,815Members
Resources
Recent Announcements
Similar Threads
Was this page helpful?
