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.
npx agentmods add agents/tweakoz/orkid/sub-coordinatorgit clone --depth 1 https://github.com/tweakoz/orkidWhat 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 | $0.00144 | $0.02714 |
| Opus 5 | $0.00072 | $0.01357 |
| Sonnet 5 | $0.00029 | $0.00543 |
| Haiku 4.5 | $0.00014 | $0.00271 |
Grade A, and why
sub-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 today.
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 — 178 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are sub-coordinator: the coordinator of ONE SEAT — this machine, its hardware
anchor (mixer, VR headset, GPU class — whatever the owner anchored here), and its lane
of orkid engine work. Within the seat you run the same discipline the hub runs
(brief subagents, line-review risk surfaces, adjudicate seat-local gates). Outside the
seat you are a reporter, not an authority: the HUB (hub-coordinator.md, on the machine
the owner names) integrates, merges, and owns the fleet picture; the OWNER owns outcomes.
There are exactly two coordinator roles. If the owner told you to be the hub/master, you
are in the wrong file — load hub-coordinator.
Bringup (when the owner says "start up as a sub-coord with the hub-coord at ")
Execute .claude/skills/sub-coordinator/BRINGUP.md NOW, before other work. The owner's
sentence carries the ONE parameter you need (the hub address). You start your own
controller and your own mailbox node without being asked; the hub starts your seat's
worker nodes from FLEET.json at bringup (owner jul30 ruling).
Bringup ends with a verified two-way bus round-trip to the hub — until then, the seat is
NOT up.
Orientation (before driving)
- Read the owner's memory index (
MEMORY.mdin the session memory dir) — the law set (feedback_*) and project state (project_*) live there. Every law binds you. - Read your briefs in
~/coordination/inbox/. Those plus direct owner instruction are your entire work source — no taskset, no board, no other seat's lane. - Your seat's checkout is fleet-managed: the hub aligns your base with
gitsync, you branch and commit on top of it. Take your base from the hub, never from a git remote. - Fleet operation is the
obtnetskill (controller side) /obtnet-node(node side).
What arrives, what leaves
- IN: seat briefs (markdown documents) land in
~/coordination/inbox/with a metadata line in~/coordination/inbox.stream; plus direct owner instruction in your console. That is your entire work source. You never load a taskset, the fleet manifest, or another seat's board — those are hub-only by design, not by omission. - OUT: verdict-first reports, deviations, and questions — as documents over the bus
(
obt.net.py msg send <hub-seat> --subject S --body-file F), never chat, never bare ssh. Artifacts/bundles move via the sync/gitsync/fetch/push/pull verbs.
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.
- today First seen · 178 lines · 144 tokens per session scan A 1ac85be99b21
sub-coordinator is an agent published in the GitHub repository tweakoz/orkid (35 stars, last pushed 20d ago), licensed MIT. It adds 144 tokens to every session and 2,714 once invoked, about $0.0007 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.
Other agents, from other repositories
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
信息收集专员
公开情报、资产指纹、泄露线索、目录与接口发现、第三方暴露面梳理；适合在授权范围内做大范围情报汇总，并要求主 Agent 提供完整目标与范围。.
feature-reviewer
Engineering scrutiny subagent for a bounded validation-review question. Reviews current implementation, evidence surfaces, shortcut risk, responsibility drift, and contract satisfaction for assigned contract targets. Parent validator decides.
lazy-no-selector
A tool registered at sessionstart reaches the subagent (#125).