Agent tools from many descriptions in a single integration. Authorize operations, not just tool calls, and approve them when a human is required. The agent gets its own identity and never holds your credentials.
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add thingctx -- uvx thingctx '<TDS>'~/.codex/config.toml
[mcp_servers.thingctx]
command = "uvx"
args = ["thingctx", "<TDS>"]
# codex mcp add thingctx -- uvx thingctx '<TDS>'~/.cursor/mcp.json
{
"mcpServers": {
"thingctx": {
"command": "uvx",
"args": [
"thingctx",
"<TDS>"
]
}
}
}.vscode/mcp.json
{
"servers": {
"thingctx": {
"type": "stdio",
"command": "uvx",
"args": [
"thingctx",
"<TDS>"
]
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"thingctx": {
"command": "uvx",
"args": [
"thingctx",
"<TDS>"
]
}
}
}claude_desktop_config.json
{
"mcpServers": {
"thingctx": {
"command": "uvx",
"args": [
"thingctx",
"<TDS>"
]
}
}
}.trae/mcp.json
{
"mcpServers": {
"thingctx": {
"command": "uvx",
"args": [
"thingctx",
"<TDS>"
]
}
}
}