orchestrate
Skill0 starsChanged 36 days agoApache-2.0View on GitHub
Drain shadow's work queue ONCE — decompose ready goals, dispatch ready leaf tasks, review landed PRs, auto-merge green/low-risk, escalate the rest. Stateless; all state lives in the queue (ledger or Jira).
173 other projects use a version of this.
Adopted by
Most recently taken up by
- chengzi-io/piertheir own version, first seen Nov 16, 2026
- endqwerty/endstacktheir own version, first seen Nov 16, 2026
- ddivyanshsharmaa/fable-toolkittheir own version, first seen Nov 16, 2026
- ronedgecomb/anythingtheir own version, first seen Nov 16, 2026
- MithunWijayasiri/dotclaudetheir own version, first seen Nov 16, 2026
- nickthelegend/nick-skillstheir own version, first seen Nov 15, 2026
Used with
Projects with it also use define-requirements, orchestrate-plan and 9 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
sson0-erSkilldefine-requirements
Work with the user, one question at a time, to complete the requirements document (docs/claudex/<slug>/01-requirements.md). Run by the orchestrator (parent). Stays out of design and implementation.
Also in 3 projects that use this
EternallLightSkillorchestrate-plan
Use when executing a locked, reviewed PLAN.md by delegating the implementation to visible worker agents in Orca, herdr, or cmux ONE PHASE at a time — splitting the plan into context-window-sized phases, giving every phase a pristine worker context, and (when the plan touches INDEPENDENT repos) running one worker per repo IN PARALLEL. Triggers on "run the plan with a worker", "execute the plan in the other pane", "split the plan into phases and build it with workers", "build the backend and frontend in parallel panes". Not for edits you can finish inline; outside Orca, herdr, or cmux use tgc-skills:orchestrate.
Also in 3 projects that use this
sson0-erSkillrecord-evidence
How to record external facts (docs, command output, codebase findings) as evidence and design choices as decisions. Used by the researcher, designer and reviewer agents.
Also in 3 projects that use this
PinkneyJiangSkillimprove-architecture
扫描代码库中的架构摩擦点,识别浅模块并挖掘深模块演进机会。生成可视化交互式 HTML 诊断报告,并在用户确认后闭环实施深度重构。
Also in 4 projects that use this
kjnezSkillticket
Work on a JIRA/Linear ticket end-to-end: read ticket details, explore codebase, create a branch, implement with TDD, run quality checks, update ticket status, and create a PR. Use when the user provides a ticket ID to implement. Triggers on requests like "/ticket PROJ-123", "work on this ticket", "implement PROJ-123".
Also in 4 projects that use this
xfstudioSkillreverse-engineer
Expert reverse engineer specializing in binary analysis,
Also in 3 projects that use this
Other versions173 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 173
- Significant-Gravitas/AutoGPT★ 188ksame name
- expo/exact★ 4.3ksame name
- rohitg00/pro-workflow★ 2.9ksame name
- apmantza/pi-lens★ 521same name
- boshu2/agentops★ 454same name
- quantumaikr/quant.cpp★ 403same name
- Noizefield/audio-plugin-coder★ 334same name
- kagenti/kagenti★ 303same name
- varun29ankuS/shodh-memory★ 294same name
- provencher/codex-skills★ 271same name
163 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository21 other files it gives its agents.
- browserbaseSkill
- cairnSkill
- developmentSkill
- dispatchSkill
- house-styleSkill
- kimi-subagentSkill
- lean-design-standardSkill
- mergeSkill
- nexus-authSkill
- nexus-jiraSkill
- orchestratorSkill
- planningSkill
- releaseSkill
- remote-opsSkill
- reviewSkill
- securitySkill
- specSkill
- ticket-pipelineSkill
- vision-reviewSkill
- workflow-basicsSkill
- writing-skillsSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 3 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 155 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Very specific: line 29
MUST also carry shadow-queue (so they re-enter the queue next drain).
- Very specific: line 34
Snapshot the ready set: Call the queue's ready-list (per the binding) for shadow's queue.
- Distinctive: line 62
AUTO-MERGE: only when ALL hold: CI green · single-ticket scope · your review found no blocking issue · NOT cross-cutting (no deploy, …
- Distinctive: line 77
Exit: when the snapshot is handled, OR if you hit a rate-limit / repeated error (stop cleanly — the next heartbeat resumes.
- Distinctive: line 84
When in doubt about a merge, ESCALATE — never merge on uncertainty.
“Common” is measured: kept by at least 6 in 10 of the 155 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
orchestrate — one drain of shadow's autonomous loop (NEX-642)
You are shadow, woken to drain shadow's work queue ONCE, then exit. You hold no memory of prior drains — all state is in the queue + git. Re-read truth; never assume.
Queue binding
This drain targets ONE queue, named in your wake prompt. Today shadow's queue is Jira/NEX (the ledger queue is dogfood-stage). Map every queue operation below to the bound queue's tools:
| Operation | Jira/NEX (current) | ledger (future) |
| Snapshot the ready set | jira_list_ready (Ready/To-Do, shadow's queue) | ListReady |
| Create a decomposed child | jira_create issue_type=Task/Story, parent=<epic-key>, labels=["shadow-queue"] | create with parent_key |
| Mark claimed / in progress | jira_update_status → In Progress | transition In Progress |
| Park / escalate | jira_update_status → Blocked | transition Blocked |
| Close | jira_update_status → Done | transition Done |
| Read a unit | jira_get | get by key |
Queue membership (Jira): shadow only acts on issues carrying the
shadow-queue label — that is shadow's queue. Children you create on decompose
MUST also carry shadow-queue (so they re-enter the queue next drain). Leave
unlabelled backlog alone.
Procedure
- Snapshot the ready set. Call the queue's ready-list (per the binding) for
shadow's queue. On ledger this set is already skill/category/dependency-aware
(NEX-645/646); on Jira it is Ready/To-Do
shadow-queueissues, so YOU apply the readiness judgement (skip anything Blocked, missing a clear DoD, or already claimed/In-Progress-without-a-landed-PR). Treat the snapshot as fixed for this drain; newly-created/changed issues are handled the NEXT drain. - For each ready unit, classify and act (one unit at a time):
- Goal / epic (no dispatchable leaf children yet) → DECOMPOSE: break it into
leaf sub-issues, each created in the queue (per the binding: Jira
jira_createparent=<epic>+labels=["shadow-queue"]), with a clearsummary, adefinition_of_done, andskillstags for routing. Transition the goal toIn Progress. Do NOT dispatch the children this drain — they enter the ready set and are picked up next wake. - Leaf task (ready, dispatchable) → DISPATCH to a builder via the dispatch
skill (
!dispatch <builder>%<provider> repo=<r> ticket=<queue-key> …), using itsskillsto pick the builder. VERIFY ACCEPTANCE — the broker log showsbuilder job created+Submit returned err=<nil>+ the pod Running (NOT the send_chat "ok"). On confirmed acceptance, immediately transition the unit toIn Progress(claimed) so it leaves the ready set — THIS is the double-dispatch guard. If acceptance fails → escalate (see Gates).
- Goal / epic (no dispatchable leaf children yet) → DECOMPOSE: break it into
leaf sub-issues, each created in the queue (per the binding: Jira
- Reconcile dispatched units (already
In Progresswith a builder run): check whether their PR has landed (gh / the run record). If a PR is up → REVIEW it. (Interim until NEX-655: you track run→PR state yourself via the broker run log / gh; once 655 lands, read the lifecycle from ledger.) - Review → merge-or-escalate (the Gates):
- AUTO-MERGE only when ALL hold: CI green · single-ticket scope · your review
found no blocking issue · NOT cross-cutting (no deploy, proto/contract,
auth/identity, multi-repo, or scope change). Then squash-merge, delete branch,
transition the unit to
Done. - ESCALATE + PARK otherwise (cross-cutting / deploy / proto / auth / scope /
CI-red / review-found issue / ANY doubt): leave the PR open, transition the unit
to
Blocked, log a distinct lineorchestrate: ESCALATION <key> <reason>, and ping the operator (comms). Do NOT merge. Do NOT retry — it waits for the operator. Deploys ALWAYS escalate. - Builder failed/stalled (run failed/stalled per NEX-653/654): transition the unit back to ready and redispatch-with-feedback ONCE; on a second failure, escalate.
- AUTO-MERGE only when ALL hold: CI green · single-ticket scope · your review
found no blocking issue · NOT cross-cutting (no deploy, proto/contract,
auth/identity, multi-repo, or scope change). Then squash-merge, delete branch,
transition the unit to
- Groom (cheap, optional): close units whose PRs already merged. Nothing else.
- Exit when the snapshot is handled, OR if you hit a rate-limit / repeated error (stop cleanly — the next heartbeat resumes; partial progress is durable in the queue). Report a one-line summary of what you did this drain.
Hard rules
- One ticket per builder; builders run in parallel; never bundle tickets in a dispatch.
- Transition-on-dispatch is mandatory (the double-dispatch guard).
- When in doubt about a merge, ESCALATE — never merge on uncertainty.
- You are stateless: if something isn't in the queue / git / the run log, it didn't happen.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

