YangWX/crossdesk
1
0
Fork 0
mirror of https://github.com/kunkundi/crossdesk.git synced 2026-11-04 21:12:42 +08:00
Code Issues Packages Projects Releases Wiki Activity

[feat] add CrossDesk MCP with remote desktop panels

Browse Source
This commit is contained in:
dijunkun
2026-11-03 02:44:55 +08:00
parent 2fab61871e
commit f57b4cc4e3
40 changed files with 6325 additions and 28 deletions
Download Patch File Download Diff File Expand all files Collapse all files
apps/cli/main.cpp
+24 -8
View File
@@ -1,4 +1,5 @@
#include "terminal_io.h"
#include "agent.h"
#include <algorithm>
#include <cerrno>
#include <clocale>
@@ -1034,6 +1035,7 @@ const char *help = R"(crossdesk-cli <9-digit-device-id> [--renderer auto|ansi|ki
crossdesk-cli view <9-digit-device-id> [options] (compatibility alias)
crossdesk-cli doctor [--json]
crossdesk-cli bridge --stdio [--password-fd 3] [--server host] [--port 9099]
crossdesk-cli agent --stdio [--password-fd 3] [--video-ws] [--server host] [--port 9099]
Connect by device ID: read-only and silent by default; hidden password prompt on the local terminal.
Read-only: c enables control; q or Ctrl+C quits; ? shows local commands.
@@ -1053,6 +1055,7 @@ Auto rendering prefers Kitty, then Sixel, then ANSI. Explicit pixel renderers re
Explicit --fps keeps a fixed preview request. Kitty transport fallback keeps the local target unchanged.
Preview requests above 15 use the negotiated interactive 30 FPS stream; lower requests use 1/8/15 FPS.
bridge: private NDJSON v1, default 8 FPS, credentials on a separate inherited descriptor.
agent: crossdesk-automation/1, structured actions and PNG screenshots; no terminal required.
No IP direct connection, audio playback, clipboard or file transfer support.
)";
} // namespace
@@ -1087,12 +1090,17 @@ int main(int argc, char **argv) {
{"kitty", "probe_on_view"},
{"sixel", "probe_on_view"},
{"audio_playback", false},
{"bridge_version", 1}};
{"bridge_version", 1},
{"automation_protocol", "crossdesk-automation/1"},
{"automation_video", "h264-annexb/1"},
{"automation_video_transport", "loopback-websocket/1"},
{"automation_video_protocols", {"h264-annexb/1", "nv12/1"}},
{"automation_video_decoders", {"auto", "software"}}};
std::cout << d.dump(argc == 3 ? -1 : 2) << "\n";
return 0;
}
int i = 2;
if (command != "view" && command != "bridge") {
if (command != "view" && command != "bridge" && command != "agent") {
if (command.find_first_not_of("0123456789") != std::string::npos)
throw std::invalid_argument(
"expected a 9-digit device ID or a command; use --help");
@@ -1105,6 +1113,8 @@ int main(int argc, char **argv) {
double zoom = 1, cell_aspect = 0;
bool control = false, stdio = false, native = false, fps_explicit = false;
int secret_fd = 3;
bool video_websocket = false;
if (command == "agent") options.fps = 30;
if (command == "view") {
if (i >= argc)
throw std::invalid_argument("device ID required");
@@ -1124,8 +1134,10 @@ int main(int argc, char **argv) {
std::string a = argv[i];
if (a == "--control" && command == "view")
control = true;
else if (a == "--stdio" && command == "bridge")
else if (a == "--stdio" && command != "view")
stdio = true;
else if (a == "--video-ws" && command == "agent")
video_websocket = true;
else {
if (i + 1 >= argc)
throw std::invalid_argument("option value missing");
@@ -1134,6 +1146,8 @@ int main(int argc, char **argv) {
options.host = value;
else if (a == "--port")
options.port = number(value, 1, 65535);
else if (a == "--log-dir" && command == "agent")
options.log_directory = value;
else if (a == "--fps" && command == "view") {
options.fps = number(value, 1, 30);
fps_explicit = true;
@@ -1158,7 +1172,7 @@ int main(int argc, char **argv) {
} else if (a == "--zoom" && command == "view") {
native = value == "native";
zoom = native ? 1 : number(value, 1, 32);
} else if (a == "--password-fd" && command == "bridge")
} else if (a == "--password-fd" && command != "view")
secret_fd = number(value, 3, 1024);
else
throw std::invalid_argument("unsupported option");
@@ -1167,14 +1181,14 @@ int main(int argc, char **argv) {
if (renderer != "auto" && renderer != "ansi" && renderer != "kitty" &&
renderer != "sixel")
throw std::invalid_argument("unknown renderer");
if (command == "bridge" && !stdio)
throw std::invalid_argument("bridge requires --stdio");
if (command != "view" && !stdio)
throw std::invalid_argument("automation and bridge require --stdio");
Descriptor output(Duplicate(1));
if (output.value < 0)
throw std::runtime_error("cannot preserve output channel");
// All library stdout writes go to diagnostics. Only Output uses the saved
// fd.
options.log_directory = Logs();
if (options.log_directory.empty()) options.log_directory = Logs();
if (command == "view") {
Descriptor logs(OpenLog(options.log_directory + "/client.log"));
if (logs.value < 0 || Redirect(logs.value, 2) < 0)
@@ -1193,7 +1207,9 @@ int main(int argc, char **argv) {
return command == "view" ? View(options, device, renderer, control,
output.value, detail, zoom, native,
cell_aspect, fps_explicit)
: Bridge(output.value, options, secret_fd);
: command == "agent"
? Agent(output.value, options, secret_fd, video_websocket)
: Bridge(output.value, options, secret_fd);
} catch (const std::exception &e) {
if (command == "view") {
std::string msg = "crossdesk-cli: " + SafeText(e.what()) + "\r\n";