agentleFS
xAIKnown publisherxai-org/customer-education-workshops/ledgerly/.cursor/skills/dispatch-subagents/SKILL.md

dispatch-subagents

Skill3 starsChanged 6 days agoApache-2.0View on GitHub

Launches isolated Task subagents in one parallel turn for independent Ledgerly work. Use for /multitask, when the user asks to run workers at the same time, or when naming api-instrumenter or ledgerly-reviewer.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    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
  1. Distinctive:

    Do not queue four jobs in one agent.

    line 15
  2. Very specific:

    After the diffs land, launch ledgerly-reviewer on the combined diff.

    line 22
  3. Distinctive:

    No sibling cross-talk. A planner or parent only sees its own children.

    line 27
  4. Very specific:

    Do not call move_agent_to_root or SetActiveBranch for a worker worktree or branch.

    line 36
  5. Distinctive:

    Read worker diffs from the linked worktree paths.

    line 37

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “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

Dispatch subagents

Hands over breadth. Independent pieces run at once; dependent steps stay in order. Each subagent starts with clean context and cannot see the prior chat.

When to use

  • The request has two or more pieces that do not depend on each other's diffs.
  • The user said /multitask, "in parallel", or named api-instrumenter.

Do not queue four jobs in one agent. Do not start a sibling after another sibling "so it has context" — the dispatch prompt carries everything.

How to launch

  1. Find what is independent. Dependent work waits.
  2. In one turn, launch one Task subagent per piece. For route instrumentation, use api-instrumenter and name one route in that prompt.
  3. Each dispatch prompt includes: the files, the shared helper path, the constraint (no prices, no seed, no suggested-credit client, no tests/suggested-credit-api.test.ts), and how to verify that piece.
  4. After the diffs land, launch ledgerly-reviewer on the combined diff.
  5. Tell the user there is one review per task.

Constraints

  • No sibling cross-talk. A planner or parent only sees its own children.
  • Five subagents cost roughly five times one agent. Launch only what is independent.
  • You still review. Do not merge or commit unless asked.

Stay on the main worktree

Parallel workers still get their own worktrees. The coordinator chat does not move with them.

  • The parent chat stays in the primary checkout on main. Do not git checkout a worker branch there.
  • Do not call move_agent_to_root or SetActiveBranch for a worker worktree or branch. If a tool would switch this chat’s root, skip it and keep the main path.
  • Read worker diffs from the linked worktree paths. Apply them later, from main, in the plan’s order.

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.