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.
git clone --depth 1 https://github.com/ngocsangyem/MeowKitWrote 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.
[](https://agentmods.dev/commands/ngocsangyem/meowkit/orchestrate)<a href="https://agentmods.dev/commands/ngocsangyem/meowkit/orchestrate"><img src="https://agentmods.dev/badge/commands/ngocsangyem/meowkit/orchestrate/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/ngocsangyem/meowkit/orchestrate"><img src="https://agentmods.dev/badge/commands/ngocsangyem/meowkit/orchestrate.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.00480 |
| Opus 5.5 | $0.00000 | $0.00192 |
| Sonnet 5.5 | $0.00000 | $0.00096 |
| Haiku 4.5 | $0.00000 | $0.00048 |
Grade A, and why
orchestrate 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/orchestrate — Multi-Runtime Job Coordinator
Usage
/orchestrate [task description]
/orchestrate [task description] --internal
/orchestrate [job-spec.yaml]
/orchestrate [job-spec.yaml] --yes
/orchestrate --resume tasks/reports/orchestrate-<timestamp>
Purpose
Coordinate staged or parallel jobs across live-verified coding-agent runtimes
(Claude Code, Codex, Cursor, and other installed CLIs) and in-session MeowKit
subagents. Each job is routed by capability and risk, isolated in a git worktree
when it writes in parallel, captured, resumable, and blocked until an independent
arbiter verifies the results. The authoritative procedure lives in mk:orchestrate.
Dispatch
Activate mk:orchestrate with the user's task description, job-spec path, or
--resume <run-dir>. Resolve every runtime, model, and agent from live evidence
during the run — never from memory or a prior report.
Flags
--internal— Routing preference: consider in-session subagents first for jobs without an explicitruntime:. Not a hard mode — a CLI fallback is allowed when a job needs a separately selectable model or stronger isolation.--yes— Pre-approve the exact destructive scope described in the spec. Does not approve anything the spec does not already describe.
Safety notes
- Secrets are refused at intake and redacted from every artifact
(
injection-rules.md). - Destructive, deploy, release, delete, or credentialed jobs require explicit
human approval for that exact scope, recorded per
intervention-recording-rules.md.--internalprompt-only isolation may never carry destructive/credentialed work. - In-session parallel fan-out honors
parallel-execution-rules.md(max 3 concurrent agents, zero file-ownership overlap, worktree isolation for writers). - Gate 1 and Gate 2 are never self-approved by orchestration (
gate-rules.md). - Prefer a direct single-agent workflow (
mk:cook) when orchestration adds no parallelism, staged dependency, runtime diversity, or arbiter value.
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.
- 1mo ago First seen · 48 lines · 0 tokens per session scan A e3832b29409c
orchestrate is a command published in the GitHub repository ngocsangyem/MeowKit (15 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 480 tokens. 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 2026-09-03.
Other commands, from other repositories
runs
List live and persisted CCAM-launched Claude Code and Codex runs.
project-health
Audit and manage project configuration — permissions, context, docs.
tasks
../../vertical-plugins/scenarios/commands/tasks.md.
finish-tasks
Finalize a work item after all tasks are done — cleanup (currently closing the ticket) and mark the work complete.
stat
A project-status command for an embedded-software project. It reads saved project notes and shows the current state, progress, next action, or a fuller project overview.
rec
A project-recovery command that reads a small set of status files and shows the current work and next action.