agentleFS
Sign inStart free
QuixiAIQuixiAI/agi-memory/skills/installed/delegation/SKILL.md

delegation

Skill617 starsChanged 10 months agoMITView on GitHub

Hand a self-contained piece of work to a sub-agent and collect the result — parallel work, long jobs, or anything that would flood this conversation

We split this skill into the rules it gives an agent

1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write the goal the way you would brief a competent stranger: they are capable and they know nothing about this conversation.

    line 43

“Common” is measured: kept by at least 6 in 10 of the 9 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

Delegation

Spawn a sub-agent to carry out a scoped task in its own session, then collect what it found. The child gets a fresh context; only its summary comes back here.

When to Use

  • Work that would flood this conversation. Reading twenty files to answer one question is the child's job, not this turn's.
  • Genuinely parallel work. Several independent questions can run at once instead of in sequence.
  • Long-running work during a heartbeat. Spawn, let the beat end, collect on a later beat with action: "get".

When Not to Use

  • The task needs the thread of this conversation — a child cannot see it.
  • One tool call would do. Delegation costs a whole model run; do not spend it on something a single recall answers.

How

  • manage_sessions with action: "spawn" and a self-contained goal. The child inherits your tools but none of your history, so state everything it needs: what to produce, what "done" looks like, and any facts it cannot look up.
  • action: "list" and action: "get" to check on and collect results — free.
  • action: "cancel" if it is no longer needed. Do not leave work running that nobody will read.

Notes

  • Write the goal the way you would brief a competent stranger: they are capable and they know nothing about this conversation.
  • The child's intermediate reasoning is not returned. If you need the workings and not just the answer, ask for them in the goal.

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.

Check your setup, free