open-orchestrator

open-orchestrator is a skill for Claude Code, Codex from gitpcl/openorchestrator. It costs 608 tokens per session (4,509 once invoked), scanned A, original, MIT.

A keyboard-driven control panel for supervising several AI coding tools working in separate Git worktrees. A worktree is an isolated working copy of a repository.

In plain words
What is it for?
Use it to start, attach to, fix, merge, and ship work across sessions running Claude Code, Pi, Droid, OpenCode, or another coding tool.
Why use it?
It gives you one place to monitor parallel coding sessions and warns when two agents begin editing the same files.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/gitpcl/openorchestrator/open-orchestrator
Any agent
npx skills add gitpcl/openorchestrator --skill open-orchestrator
Clone the repo
git clone --depth 1 https://github.com/gitpcl/openorchestrator

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for open-orchestrator

README.md
[![agentmods](https://agentmods.dev/badge/skills/gitpcl/openorchestrator/open-orchestrator.svg)](https://agentmods.dev/skills/gitpcl/openorchestrator/open-orchestrator)
Your own site
<a href="https://agentmods.dev/skills/gitpcl/openorchestrator/open-orchestrator"><img src="https://agentmods.dev/badge/skills/gitpcl/openorchestrator/open-orchestrator.svg" alt="Measured on agentmods" height="20"></a>
Per session 608 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,509 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00608 $0.04509
Opus 5 $0.00304 $0.02254
Sonnet 5 $0.00122 $0.00902
Haiku 4.5 $0.00061 $0.00451

Measured 4d ago against content hash fefaee976572, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

open-orchestrator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

src/open_orchestrator/skills/open-orchestrator/SKILL.md · 300 lines

How it starts

The opening of the file, as written. The whole thing — 300 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Open Orchestrator — the multi-provider cockpit for parallel AI coding

Open Orchestrator (owt) supervises parallel AI coding sessions across isolated git worktrees. It does not try to be the agent — it hosts whatever AI coding tool you point it at (Claude Code, Pi, Droid, OpenCode, or a custom tool) and gives you one place to watch and steer them all.

Two things make it worth using, and native tooling lacks both:

  1. A persistent cross-worktree control plane — a standing board across long-lived worktrees, three lanes deep (NEEDS YOU / READY TO SHIP / IN FLIGHT), driven entirely from the keyboard.
  2. Conflict Guard — real-time file-overlap detection that warns you the moment two agents start editing the same files, long before they collide at merge.

The primary interface is one command: owt — it launches the control plane. Press n to start work (type a task, pick one worktree or a native plan-first workflow, confirm), a to attach, s to ship, f to fix conflicts, m to merge. Every row carries verb actions and the footer shows only the keys that apply to the focused row.

You own the cockpit; the AI tools own the engine.

Commands (scripting / CI reference)

Humans drive the control plane with owt and the keys above; the verbs below are the same actions exposed for scripts, pipelines, and automation.

Command Alias Description
owt Launch the Control Plane — prioritized lanes, row verbs
owt new "task" owt n Create worktree + session + deps + AI agent. One command.
owt new "task" --workflow Launch a native plan-first Claude Code workflow in the worktree
owt new "task" --ai-tool <name> Pick the provider (claude/pi/droid/opencode/custom)
owt new "task" --herdr Use the herdr multiplexer backend instead of tmux
owt new "task" --tmux Force tmux backend (default; useful to override config)
owt new "task" --headless Create worktree without tmux (CI/script use)
owt branch "task" Create branch in current checkout instead of worktree (faster, zero disk)
owt list owt ls Quick text list of worktrees (non-interactive, for scripts/pipes)
owt switch <name> owt s Jump to a worktree's session
owt attach <name> Hand off to the worktree's session via the active backend (--herdr / --tmux to override)
owt send <name> "msg" Send command to a worktree's AI agent
owt send --all "msg" Broadcast to ALL worktrees
owt send --working "msg" Broadcast to WORKING worktrees only
owt merge <name> owt m Two-phase merge + Conflict Guard + auto-cleanup (--rebase, --strategy, --leave-conflicts)
owt ship <name> Commit + merge + delete in one shot
owt delete <name> owt rm Delete worktree + session + status
owt queue Show optimal merge order (with overlap counts) for completed worktrees
owt queue --ship Ship all completed worktrees in optimal order
owt wait <name> Poll until agent finishes (for CI/scripts)
owt note "msg" Share context across all agent sessions
owt sync [--all] Sync worktree(s) with upstream
owt cleanup [--force] Remove stale worktrees (dry-run by default)
owt config validate Validate configuration file
owt config show Display effective config as TOML
owt db purge [--days N] Delete messages older than N days (default 30)
owt db vacuum Optimize and compact the database
owt db health [--check] Database health diagnostics with CI thresholds
owt doctor [--fix] Diagnose and fix orphaned resources
owt usage [--days N] Local usage counts (cockpit launches, worktrees started)
owt --theme <name> Override UI theme (auto, dark, light, dark-ansi, light-ansi)
owt --json <cmd> Machine-readable JSON output for list, queue, doctor, db health
owt version Show version

Read the full file on GitHub · 300 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago First seen · 300 lines · 608 tokens per session scan A fefaee976572

Subscribe to this mod's changes

open-orchestrator is a skill published in the GitHub repository gitpcl/openorchestrator (5 stars, last pushed 1mo ago), licensed MIT. It adds 608 tokens to every session and 4,509 once invoked, about $0.0030 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

flow

Orchestrator for Ticket-Flow — default is --local (all phases in this session with user checkpoints). --parallel works multiple ready tickets at once via worktree-isolated subagents in this session; --serial (one subagent at a time, merge+deploy+cleanup per ticket) and --loop (re-query the ready queue after every…

keefar/ticket-flow · 292 tokens

task-ledger

Ledger development tasks to Feishu Base via dev-task-ledger.sh. Write roots from /.openclaw/tasks-allowlist.json (per agent). Use on accept/confirm/close of coding/ops work; not for pure Q&A.

yunsii/wezdeck · 53 tokens

init

Scaffold a ticket-flow project — one-time setup; trigger on "setz ticket-flow auf", "richte das Projekt für ticket-flow ein", "set up ticket-flow here". Runs bd init with tf's agents template (keeps Claude Code's own memory system), scaffolds spec dirs and worktree settings, migrates an existing KANBAN.md into beads.…

keefar/ticket-flow · 108 tokens

pickup

Internal phase 1 of ticket-flow, normally invoked by ticket-flow:flow — validate Definition of Ready, create (or adopt) the ticket's isolated worktree, set the branch lock, claim the item atomically → In Progress. Invoke directly only for recovery, or when the user explicitly wants just the claim+worktree step. Args…

keefar/ticket-flow · 109 tokens

implement

Internal phase 2 of ticket-flow, normally invoked by ticket-flow:flow — execute the plan for the claimed ticket inside its worktree: incremental commits, typecheck/test after each major step. Invoke directly only to continue implementing a ticket that is already picked up.

keefar/ticket-flow · 55 tokens

kanban

Use when a prompt contains a new bug/feature/change not yet tracked, or when an item's status changes — capture and maintain it in the project's beads tracker (Inbox→Backlog with Definition of Ready, priorities, the pipe-separated note format).

keefar/ticket-flow · 52 tokens