cmuxlayer
cmuxLayer
Deine KI-Agenten können die Terminals der jeweils anderen nicht sehen. Einer läuft in Tab 1, ein anderer in Tab 2 – und du bist die Zwischenablage zwischen ihnen. cmuxLayer löst das: 26 MCP-Tools, die KI-Agenten die programmatische Kontrolle über Terminal-Arbeitsbereiche geben.
Schnellstart
npm install -g cmuxlayerErfordert, dass cmux läuft.
Füge dies zu deiner MCP-Konfiguration hinzu:
Codex CLI / T3 Code
T3 Code übernimmt MCP-Server aus der Codex CLI-Konfigurationsdatei unter ~/.codex/config.toml (oder $CODEX_HOME/config.toml).
[mcp_servers.cmux]
command = "cmuxlayer"Claude Code, Cursor, VS Code, Claude Desktop
{
"mcpServers": {
"cmux": {
"command": "cmuxlayer"
}
}
}Konfigurationspfade: Codex CLI / T3 Code
~/.codex/config.toml(oder$CODEX_HOME/config.toml) | Claude Code.mcp.jsonoderclaude mcp add cmuxlayer -s user -- cmuxlayer| Cursor.cursor/mcp.json| VS Code.vscode/mcp.json| Claude Desktop — siehe MCP-Dokumentation für plattformspezifische Pfade
Related MCP server: hyperpanes-mcp
Was du tun kannst
Sage deinem KI-Agenten Dinge wie:
"Teile einen Bereich nach rechts und führe dort meine Test-Suite aus"
"Starte einen Claude Code-Agenten in einem neuen Bereich, um auth.ts zu refactoren"
"Lies den Bildschirm von surface:2 und sag mir, ob der Build erfolgreich war"
"Warte, bis alle Agenten fertig sind, und lies dann ihre Ausgabe"
"Setze den Status der Seitenleiste, um unseren Deployment-Fortschritt anzuzeigen"
Unter der Haube stellt cmuxLayer 26 MCP-Tools für Terminalsteuerung, Bildschirmlesen, Layout-Management und Multi-Agenten-Orchestrierung bereit. read_screen analysiert Agenten-Metadaten (Status, Modell, Token, Kontext-%) für Claude Code, Codex, Gemini und Cursor.
MCP-Tools (26)
Alle Tools werden mit ToolAnnotations für die automatische Durchsetzung von Sicherheitsrichtlinien ausgeliefert.
Terminalsteuerung — new_split new_surface move_surface reorder_surface send_input send_key read_screen rename_tab close_surface browser_surface
Agenten-Lebenszyklus — spawn_agent send_to send_to_agent wait_for wait_for_all interact stop_agent kill
Arbeitsbereich — list_surfaces list_agents my_agents get_agent_state read_agent_output notify set_status set_progress
Nur Lesen (6)
Tool | Was es tut |
| Listet alle Oberflächen über Arbeitsbereiche hinweg auf |
| Liest Terminalausgabe mit analysiertem Agentenstatus |
| Vollständiger Status eines verfolgten Agenten |
| Alle Agenten, mit optionalen Filtern |
| Kinder eines übergeordneten Agenten mit Live-Bildschirmstatus |
| Strukturierte Ausgabe zwischen Trennmarkierungen |
Verändernd (17)
Tool | Was es tut |
| Erstellt einen Terminal- oder Browser-Split-Bereich |
| Erstellt einen Tab in einem bestehenden Bereich |
| Verschiebt eine Oberfläche in einen anderen Bereich oder an eine andere Position |
| Sortiert Tabs innerhalb eines Bereichs neu |
| Sendet Text an eine Oberfläche |
| Sendet Tastendruck (Return, Escape, Strg-C, etc.) |
| Benennt einen Oberflächen-Tab um |
| Zeigt ein cmux-Benachrichtigungsbanner an |
| Setzt Schlüssel-Wert-Paar für den Seitenleistenstatus |
| Setzt Fortschrittsanzeige (0.0-1.0) |
| Interagiert mit Browser-Oberflächen |
| Startet einen CLI-Agenten in einem neuen Bereich |
| Sendet Text an einen verfolgten Agenten, ohne dessen Oberfläche zu kennen |
| Sendet einen Prompt an einen laufenden Agenten |
| Blockiert, bis der Agent einen Zielstatus erreicht (Standard: |
| Blockiert, bis mehrere Agenten fertig sind |
| Sendet interaktive Eingaben (bestätigen, abbrechen, fortsetzen) |
Destruktiv (3)
Tool | Was es tut |
| Schließt einen Terminal- oder Browser-Bereich |
| Stoppt einen Agenten ordnungsgemäß |
| Beendet Agentenprozesse gewaltsam |
Unterstützte Agenten
CLI | Befehl | Automatisch erkannt |
Claude Code |
| Status, Modell, Token, Kontext-% |
Codex |
| Status, Modell, Kontext-% |
Gemini CLI |
| Status, Modell, Token, Kontext-% |
Cursor |
| Status, Modell, Token, Kontext-% |
|
Architektur
AI Agent ─── MCP ───> cmuxLayer ─── Unix socket ───> cmux
├── Agent engine (spawn → monitor → teardown)
├── Screen parser (5 agent formats)
├── Mode policy (autonomous vs manual)
└── State manager + event logDer Socket-Client verbindet sich über einen Unix-Socket mit cmux. Er verbindet sich bei Trennung automatisch neu und greift auf den CLI-Subprozess zurück, falls der Socket nicht verfügbar ist.
Verbindung | Latenz | Beschleunigung |
CLI-Subprozess | ~142ms | Basislinie |
Unix-Socket | ~0.1ms | 1.423x |
Fehlerbehebung
cmux läuft nicht cmuxLayer erfordert eine laufende cmux-Instanz. Installiere es zuerst und starte dann eine cmux-Sitzung, bevor du cmuxLayer verwendest.
Tools erscheinen nicht in Codex CLI oder T3 Code
Starte den Client neu, nachdem du cmuxlayer zu ~/.codex/config.toml hinzugefügt hast. Wenn du ein benutzerdefiniertes Codex-Home verwendest, überprüfe, ob $CODEX_HOME/config.toml denselben mcp_servers.cmux-Eintrag enthält.
Tools erscheinen nicht in Claude Code
Starte Claude Code neu, nachdem du die MCP-Konfiguration hinzugefügt hast. Führe claude mcp list aus, um zu überprüfen, ob cmuxlayer verbunden ist.
Socket-Verbindung fehlgeschlagen
cmuxLayer erkennt den cmux-Socket automatisch (macOS: ~/Library/Application Support/cmux/cmux.sock). Überschreibe dies bei Bedarf mit CMUX_SOCKET_PATH.
Testen
bun run test # 406 tests via vitest
npm run typecheck # Type checkingEntwicklung
npm install
npm run dev # Run with tsx (hot reload)
npm run build # Compile TypeScript
npm start # Run compiled outputMitwirken
Siehe CONTRIBUTING.md für die Entwicklungseinrichtung und PR-Richtlinien.
Lizenz
Apache 2.0 — siehe LICENSE.
Teil des Golems KI-Agenten-Ökosystems. cmuxlayer.etanheyman.com | Erstellt von @EtanHey.
Available Tools
10 toolsclose_surfaceADestructive
Close one surface, managed agent, or workspace with live-agent guards. scope="agent" stops the agent AND closes its pane, and reports the two halves separately (agent_stopped, surface_closed) so a pane that survives is never reported as closed. The pane close obeys the same live-agent guard as scope="surface": without force:true a still-live agent keeps its pane, and the receipt says so.
Inputs: force, scope, surface, agent_id, +1 more. Returns: ok, pane, force, scope, +12 more
control_healthARead-onlyIdempotent
Report terse control-path health by default; pass detail=full for diagnostics.
Inputs: detail. Returns: ok, health, retry_count
list_agentsA
List live-derived agents, including registry-persisted prompt blockage and pause state; filter to blocked agents or children with mine/parent_agent_id. Default summary returns flat addressable scalars and hides close tombstones and failed spawns whose surfaces are absent; request a terminal state or detail=full to include them. Full detail also includes provenance, health diagnostics, the registry record, and up to 20 unresolved or attention delivery receipts.
Inputs: mine, repo, model, state, +5 more. Returns: ok, count, agents, derived_at, +1 more
list_surfacesARead-onlyIdempotent
List workspace, pane, and surface topology. Condensed by default; verbose=true adds raw cmux fields.
Inputs: verbose, workspace, preview_lines, include_screen_preview. Returns: ok, surfaces, workspaces, retry_count, +1 more
read_screenARead-onlyIdempotent
Read a terminal screen and parsed harness status. Use raw=true for full text or parsed_only=true for monitoring.
Inputs: raw, lines, surface, workspace, +3 more. Returns: ok, parsed, surface, retry_count
report_to_parentA
Raise a short blocker to this managed agent's registry parent. cmuxlayer chooses the parent; callers cannot address arbitrary agents. The blocker is durably appended to the parent's inbox and its pointer is actively delivered. If that wake fails, cmuxlayer alerts the nearest reachable ancestor and returns fallback provenance. A root agent has no parent and receives an error. Workers with collab_path must append there to reach their own parent lead; this tool refuses that upward route.
Inputs: blocker (required). Returns: ok, route, durable, delivery, +6 more
send_toA
Send text or a key through the shared delivery engine. send_to submits new messages. Key-Return supports pickers, menus, permission prompts, and recovery of this caller's exact typed draft on the same agent/session once the displayed unambiguous reset time passes or new harness output confirms recovery; the old refusal may stay visible. Exact /usage, /usage-credits or /rate-limit-options navigation is also guarded. Read the pane and receipt first; never retype or repeat accepted input. Every receipt includes caller_agent_id (null when unknown). Workers with collab_path cannot address their own parent or ancestor leads in any mode; append to that collab file instead. Unknown callers remain allowed. Lead-originated and engine-internal pushes remain allowed. Targets may be one agent, structured agent targeting, or a raw surface in surface/command/key mode; for shell commands use mode:command. Text success requires a verified submit or new matching queue proof. Otherwise ok=false with submit_unverified retains pending verification and ownership for retries without retyping. Verified text success returns only ok, target identity, delivery_id, delivery_state, submitted, caller_agent_id, and queued_behind_turn when applicable. Key success returns its key verification fields. Pass verbose=true for full diagnostics; failures always retain them.
Inputs: mode, text, target, surface, +13 more. Returns: ok, key, model, title, +37 more
spawn_agentA
Spawn a managed agent or terminal, or resume a captured agent on a fresh surface preserving its ID. Resume: {resume_agent_id:"..."}. Omit new-spawn fields: repo, cli, model, effort, cwd, prompt, boot_prompt_path, worktree, mcp_profile, role, placement, authority, collab_path, parent_agent_id, max_cost_per_agent; these cause INVALID_RESUME_SPEC. Optional resume controls such as force, workspace, focus, report_path and verbose remain supported. Resume restores session context; it does not resend a task. For a new task, check current readiness, then use send_to separately. ok:true/state:started means lifecycle started, not ready or task delivered. At a chooser or permission prompt, inspect the current frame with read_screen before acting; never blindly send Return or assume success. pending_verify leaves delivery unverified; follow wait_for guidance in warning when present; do not resend blindly. Placement is deterministic; boot_prompt_timeout_ms also bounds pane placement. Default successes have at most six fields: ok, agent_id, surface_id, state (spawn outcome), delivered (boot prompt), and optional warning. Use verbose:true for contract_path, report_path, done_marker, transport, worktree, and diagnostics; report_path and done_marker are also in the issued contract file. Failures always keep full detail.
Inputs: cli, cwd, repo, role, +25 more. Returns: ok, cwd, role, type, +26 more
update_surfaceB
Move or rename a surface, or adopt_session to rebind an existing managed Claude or Codex identity after a restart using pinned historical evidence and fresh process/topology proof. Adoption changes no pane or input; layout reconciliation follows separately.
Inputs: action (required), surface (required), cwd, pane, +20 more. Returns: ok, pane, title, action, +3 more
wait_forB
Block until one agent_id or every agent in ids reaches a target registry state and return health. Defaults to waiting for completion (done).
Inputs: ids, mine, watch, agent_id, +6 more. Returns: ok, typed, watch, results, +16 more
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 10 tools
Each tool has a reasonably distinct purpose: spawn, send, list, close, read, wait, health, report. There is mild overlap between list_agents and list_surfaces (agents live on surfaces) and among the surface-touching tools (spawn_agent, update_surface, close_surface), but descriptions clarify the boundaries.
Nearly all names are snake_case with a verb_noun or verb_prep pattern (control_health, spawn_agent, list_agents, update_surface, close_surface, list_surfaces, read_screen). Minor deviations are the prepositional forms send_to, wait_for, and report_to_parent, which are still readable and consistent in style.
Ten tools is well within the ideal range and each covers a distinct facet of agent/surface orchestration. No redundant or filler tools appear; the set feels tightly scoped to the domain.
The surface covers the full lifecycle: spawn/resume, list, send, read, wait, update/adopt, close, report, and health. Coverage is strong, with only minor gaps such as no dedicated agent-state update beyond update_surface's adoption path.
Maintenance
Related MCP Connectors
Real-time chat for AI agents. Claude Code, Cursor, Cline and Codex join channels over MCP.
Shared memory MCP server for Claude Code, Codex, Gemini CLI and Cursor. Cited, editable.
Remote MCP server exposing your Node Command fleet (files, shell, processes, browser) to AI hosts.
Real-time chat hub for AI agents — Claude Code, Cursor, Cline, Codex over MCP or REST.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceTerminal MCP server for AI coding agents with persistent PTY sessions, ring-buffer incremental reads, headless xterm screen capture, multi-agent orchestration, and a real-time web dashboard.33 npm28MIT
- AlicenseAqualityBmaintenanceMCP server for hyperpanes terminal workspace app, enabling AI agents to compose and launch workspace layouts, inspect and drive terminal panes, stream output, and orchestrate agent hierarchies.501MIT
- AlicenseAqualityBmaintenanceMCP server to control Onda terminal from AI agents, providing tools for splitting panes, running commands, managing tabs and workspaces, and orchestrating multi-agent workflows across multiple windows.395 npmMIT
- AlicenseAqualityDmaintenanceMCP server that launches and supervises coding-CLI subagents (claude, opencode) in detached tmux windows, collapsing the ceremony of starting a TUI coding assistant into a single tool call.5MIT