orchestrate
Skill0 starsChanged yesterdayMITView on GitHub
Plan, dispatch and evaluate a multi-part request through pier: a written plan with checkable acceptance criteria, waves of parallel Pi runs, and an independent evaluator run before reporting done. Use before the first pier run of a request with several deliverables or new code beyond a small edit.
173 other projects use a version of this.
Adopted by
Most recently taken up by
- 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
- YuY-QK/agent-kittheir own version, first seen Nov 15, 2026
Used with
Projects with it also use define-requirements, orchestrate-plan and 10 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.
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 155 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 18
Before the first dispatch, write the plan in the reply:
- Distinctive: line 26
Domain checklist: if an acceptance checklist exists for the domain (a skill or a file), note its path.
- Very specific: line 56
After each integration point (the skeleton, a merged wave, a finished feature), run evaluator on the integrated result.
“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
The main session plans, decides and judges; Pi runs do the legwork. This skill is the loop; the pier skill has the commands. An unplanned split serializes the work, and a criterion nobody checks ships broken.
1. Plan
Before the first dispatch, write the plan in the reply:
- Outcome and acceptance criteria. Each criterion names what to run or open, what to observe, and what passes. "The panel scrolls independently" is a wish; "with the panel full at 1280×800, the page does not scroll and the panel's last control is reachable by scrolling the panel" is a criterion.
- Cross-cutting criteria (whole-page layout, an end-to-end flow, startup) get their own lines: no single worker owns them, so only the evaluator checks them.
- Domain checklist: if an acceptance checklist exists for the domain (a skill or a file), note its path. The evaluator checks it on top of the plan's criteria, which catches what the plan left out.
- Deliverables, split by module or concern, not by stage. Shared interfaces (types, schemas, fixtures) are written to files here first; each deliverable lists the files it writes, the files it reads, and what it waits for.
- Size: one deliverable = one pier run that finishes within the run deadline.
Bigger work (a whole app, a stack of features) is split further: a deadline hit costs
a
send "continue"round. - Skeleton first: a new app, service or UI starts from the official scaffold or template (the framework's init command, an official layout block), and the skeleton passes evaluation before workers fill it in. Workers extend a checked skeleton; they do not invent one.
Tiny edits, tightly coupled changes and decisions stay here.
2. Execute
- Independent deliverables start in one tool-call block, one background
pier runeach, before waiting for any result; one agent can have several concurrent runs. Deliverables beyond the free slots (pier skill) go in the next wave. - Writers in the same wave own disjoint files.
- A brief is self-contained: goal, paths to read, files it owns, constraints, and a
Verifylist holding every acceptance criterion of the deliverable, including those that need the running product. A requirement stated only in prose goes unchecked. - While a wave runs, write the next wave's briefs or the evaluator brief. Dependent work starts when its inputs land, with their findings in its brief.
3. Evaluate
- After each integration point (the skeleton, a merged wave, a finished feature), run
evaluatoron the integrated result. Its brief: the outcome, every acceptance criterion (deliverables plus cross-cutting), the domain checklist path, how to start and stop the product, and where to save screenshots. Leave out the workers' reports: the evaluator judges the artifact, not the claims. It can overlap the next wave only if that wave does not change what it checks. - A failed criterion goes back to the worker session that owns it:
pier send <ref> "<criterion> failed: <evidence>", then that criterion is checked again. After two failed rounds, decide here: rewrite the spec, split it, or take it over. - Spot-check one cited verdict. Judge here what a script cannot: for UI, read two or three of the evaluator's screenshots.
- Report done only with an evaluator pass, or name the criteria still failing.
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.

