Browser-based graphical shell and control plane for a Linux server you own — real terminal, file manager, live metrics and BYOK AI in one mobile-first workspace. Next.js 16 · React 19 · node-pty · Tailwind v4 · shadcn/ui. Public alpha, self-hosted.
TypeScript browser-shellcontrol-planefile-managerhomelabmobile-firstnextjsptyreactself-hostedserver-managementshadcn-uisysadmintailwindcssterminaltypescriptvpsweb-desktopweb-terminal
import { resolveMcpInstallTarget } from "@/lib/host/mcp-install-target";
import { callManagedScProvider } from "./managed-sc-call";
import { projectMcpResult, projectMcpStructuredProjection } from "@/lib/host/project-mcp-result";
import { callProjectMcpTool, listProjectMcpToolPage, publicProjectMcpServers, readProjectMcpServers } from "@/lib/host/projects-api";
import { type McpTool, S, str, mcpDirect } from "./tool-kit";
import { requireWorkflowProjectTarget } from "./workflow-workspace-guard";
export const PROJECT_MCP_TOOLS: McpTool[] = [
{
name: "project_mcp_tools",
result: { maxTextBytes: 64 * 1024, overflowHint: "Use limit=1 for a narrower descriptor page; preserve the exact nextCursor." },
title: "List Project MCP Tools",
description: "Discover tools from one MCP server declared by an explicitly selected project's .mcp.json. Server aliases and tool schemas are returned on demand; config, env, headers, and credentials are never returned and project tools never join MSO's global catalog.",
chatgptDescription: "Discover live tools of an installed MCP. project=@host selects this VPS; other values select an exact project. Find aliases with project_mcp_manage inspect. Follow nextCursor; refresh bypasses descriptor cache.",
scope: "exec",
annotations: { readOnlyHint: false, destructiveHint: true, openWorldHint: true, idempotentHint: false },
actionContract: { phase: "discover", target: "project-mcp", sourceOfTruth: "provider", discover: ["project_capabilities"], validators: ["project-selection", "private-connection"], confirmation: "none", concurrency: "provider", presentation: "structured" },
audit: { action: "exec.run" as const, targetArg: "server" },
limit: { key: "projects.mcp.read", max: 30, windowMs: 60_000 },
inputSchema: S({
project: { type: "string", description: "Exact project id/path/name from projects_list, or @host for host-installed MCPs." },
server: { type: "string", description: "MCP server alias returned by project_capabilities." },
cursor: { type: "string", maxLength: 8192, description: "Exact nextCursor from the previous page." },
limit: { type: "integer", minimum: 1, maximum: 100 }, refresh: { type: "boolean", description: "Bypass the 30-second HTTP descriptor cache." },
}, ["project", "server"]),
run: async (a) => {
const project = await resolveMcpInstallTarget(str(a, "project"));
const servers = publicProjectMcpServers(await readProjectMcpServers(project.path));
const selected = servers.find((server) => server.name === str(a, "server")); if (!selected) throw new Error("project MCP server not found");
return { project: { id: project.id, name: project.name }, server: selected, ...await listProjectMcpToolPage(project.path, selected.name, { cursor: typeof a.cursor === "string" ? a.cursor : undefined, limit: Number(a.limit) || 50, refresh: a.refresh === true }) };
},
},
{
name: "project_mcp_call",
title: "Call Project MCP Tool",
description: "Call one exact tool on one MCP server declared by a validated project's .mcp.json. The project MCP configuration stays server-side; MSO launches/calls it with credential-scrubbed process environment or a guarded remote transport and returns only the MCP tool result.",
chatgptDescription: "Call one discovered tool through MSO to an installed external MCP. Use project=@host for this VPS or an exact project. Config, credentials, and authorization stay server-side; inspect and discover first.",
scope: "exec",
outputSchema: { type: "object", properties: { result: { type: "object" } }, required: ["result"], additionalProperties: false },
annotations: { readOnlyHint: false, destructiveHint: true, openWorldHint: true, idempotentHint: false },
actionContract: { phase: "execute", target: "project-mcp", sourceOfTruth: "provider", discover: ["project_capabilities", "project_mcp_tools"], validators: ["dynamic-input-schema", "project-scope", "private-connection"], confirmation: "contextual", concurrency: "provider", presentation: "tool-owned" },
limit: { key: "projects.mcp.call", max: 30, windowMs: 60_000 },
audit: { action: "exec.run" as const, targetArg: "server" },
result: { maxTextBytes: 64 * 1024, overflowHint: "Project MCP result was compacted; request a narrower project tool call." },
inputSchema: S({
project: { type: "string", description: "Exact project id/path/name from projects_list, or @host for host-installed MCPs." },
server: { type: "string", description: "Project MCP server alias from project_capabilities." },
tool: { type: "string", description: "Exact tool name from project_mcp_tools." },
arguments: { type: "object", description: "Arguments matching that dynamic tool's input schema.", additionalProperties: true },
}, ["project", "server", "tool"]),
run: async (a, context) => {
const project = await resolveMcpInstallTarget(str(a, "project"));
if (project.installationScope === "project") await requireWorkflowProjectTarget(context, project.path);
const managed = await callManagedScProvider(project.path, str(a, "server"), str(a, "tool"), a.arguments ?? {});
const out = projectMcpResult(managed.handled ? managed.result : await callProjectMcpTool(project.path, str(a, "server"), str(a, "tool"), a.arguments ?? {}));
return mcpDirect(out.content, out.isError, { result: { project: project.id, server: str(a, "server"), tool: str(a, "tool"), output: projectMcpStructuredProjection(out) } }, out.meta);
},
},
];