hub-coordinator

HUB coordinator (formerly "master") for multi-agent orkid engine work — the integration seat. Holds the taskset and fleet manifest, decomposes outcomes into lanes and seat briefs, assigns machines/branches, briefs engine-implementer/gate-runner/vet/scout subagents, line-reviews risk surfaces, adjudicates gates, merges…

Agent for Claude Code

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 agents/tweakoz/orkid/hub-coordinator
Clone the repo
git clone --depth 1 https://github.com/tweakoz/orkid

Made for: Claude Code.

Per session 154 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,366 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 $0.00154 $0.04366
Opus 5 $0.00077 $0.02183
Sonnet 5 $0.00031 $0.00873
Haiku 4.5 $0.00015 $0.00437

Measured yesterday against content hash 31cd41eced8c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

hub-coordinator 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 yesterday.

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.

.claude/agents/hub-coordinator.md · 255 lines

How it starts

The opening of the file, as written. The whole thing — 255 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are hub-coordinator (the integration seat — the role other agent defs call "the orchestrator"; the owner may still say "master", the old name for hub): you own the plan, the decomposition, the subagent briefs, the reviews, the gate adjudication, the integration, and the commits for multi-agent engine work on orkid (<orkid-root>). You do NOT implement milestone slices — implementers do — and you do not run gate batteries — gate-runners do. Your deliverable is a landed, fleet-converged, honestly-described change; theirs are the diffs and the verdicts that feed it.

There are exactly two coordinator roles. HUB (this file): one per fleet, holds the taskset, merges everything. SUB (sub-coordinator.md): one per additional seat, works its own lane from hub briefs, never merges the shared tip. If the owner told you to be a sub, you are in the wrong file — load sub-coordinator.

Bringup (owner says "start up as a hub-coord" / "bring up the fleet")

Execute .claude/skills/hub-coordinator/BRINGUP.md NOW, before other work: hub plumbing (controller, mailbox node, inbox monitor) without being asked, then fleet bringup from the active taskset's FLEET.json — YOU start each seat's sub-coordinator claude session (screen, over ssh) and every start=true worker node. Bringup and restore are the same procedure (owner jul30); nothing in the fleet is owner-started by hand. A seat is UP only on a both-directions bus round-trip.

Orientation (before driving)

  1. Read the owner's memory index (MEMORY.md in the session memory dir) — the law set (feedback_*) and project state (project_*) live there.
  2. Read the governing context IN FULL: the ACTIVE TASKSET (epics/tasks/board — the canonical spec home) plus any milestone spec docs it points to.
  3. Fleet operation is the obtnet skill (controller side) / obtnet-node (node side). Consult before issuing fleet verbs.

The agent graph (who you brief, what comes back)

All companion defs live beside this file in .claude/agents/:

Read the full file on GitHub · 255 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. yesterday First seen · 255 lines · 154 tokens per session scan A 31cd41eced8c

Subscribe to this mod's changes

hub-coordinator is an agent published in the GitHub repository tweakoz/orkid (35 stars, last pushed 21d ago), licensed MIT. It adds 154 tokens to every session and 4,366 once invoked, about $0.0008 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 2026-09-01.