agentleFS
mblaubergmblauberg/provenant/skills/orchestrate/SKILL.md

orchestrate

Skill1 starsChanged 4 days agoMITView on GitHub

Use for template graphs, composites, parallel work, review panels and bounded repair loops. Not for single-owner work or coupled defects (diagnose). Session owns continuing goals and STOP missions.

161 other projects use a version of this.

Adopted by

Most recently taken up by

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-erSkill

    define-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

  • EternallLightSkill

    orchestrate-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-erSkill

    record-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

  • PinkneyJiangSkill

    improve-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

  • kjnezSkill

    ticket

    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

  • xfstudioSkill

    reverse-engineer

    Expert reverse engineer specializing in binary analysis,

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 144 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use task for one bounded assignment, plan when the shape is unsettled, batch for parallel implementation and deliver for implementation …

    line 8
  2. Very specific:

    Use explore over areas, research for primary sources and review for a pinned head.

    line 21
  3. Distinctive:

    Use separate worktree and write_scope bindings per slice.

    line 27
  4. Very specific:

    Use steer for a message, stop for a subtree, retry for unfinished work, and edit for a graph patch.

    line 32
  5. Distinctive:

    Verify commits, paths, checks and routes against output.

    line 45

“Common” is measured: kept by at least 6 in 10 of the 144 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Orchestrate

Use task for one bounded assignment, plan when the shape is unsettled, batch for parallel implementation and deliver for implementation through landing. A composite is a template containing nodes; dotted names select flavours such as review.security and implement.fix. Slice larger work into lanes of about three acceptance items each, merged into one integration branch.

provenant run --route junior --file brief.md --explain
provenant run --use plan --file brief.md --explain
provenant run --use batch --file brief.md --set 'worktree="/absolute/project/.worktrees/slice"' --set 'write_scope=["src"]' --explain

Use explore over areas, research for primary sources and review for a pinned head. Nodes use over, after and inputs for copies, dependencies and evidence. Layers are bundled, user, repository and inline; project templates go under .provenant/templates/. See the template runbook.

Register writer worktrees first with provenant worktree create NAME --base BASE. Use separate worktree and write_scope bindings per slice; review the integrated head. The planner brief owns slicing and tier choices; brief and collect owns the hand-off. Keep coupled writes with one owner.

Orient with provenant status; collect with provenant output TARGET. Use steer for a message, stop for a subtree, retry for unfinished work, and edit for a graph patch. Use --route junior by default for implementation slices; raise the tier for design. The receipt records the resolved adapter/model@effort. Inside a worker, run --use TEMPLATE adds a child to the same graph, sharing budget, stopping and trace. From inside a lane, status reads that graph; output CHILD and output CHILD --trace collect a child's result, failure cause and timeline through the host relay. Use IDs from the launch receipt. The lane needs no journal or credential access. A child may select another project with --project; register writer worktrees in that project first. Credential and state roots remain refused. Attached live children are cancelled when their parent settles, including a successful exit. Use --wait to collect their completion before exiting, or explicitly detach a child that should outlive its parent. Any agent may return a patch; started attempts stay immutable.

Verify commits, paths, checks and routes against output. A synthesis cannot waive a gate. Bound repair and independently review its delta. HARNESS owns review floors and authority; the chair keeps user communication and acceptance. Load routing for tier and independence decisions and writer boundaries for confinement. See Herdr for observation.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.