agentleFS
AL-JANEFAL-JANEF/janef-forge/skills/engineering/delegation-discipline/SKILL.md

delegation-discipline

Skill1 starsChanged 54 days agoMITView on GitHub

Decide whether subagents, parallel agents, or agent teams are justified and, if so, delegate safely: one agent unless work is bounded, independent, and clearly owned; explicit outcome, ownership, constraints, and return format per worker; the integrating agent owns final correctness, security review, and verification. Use whenever delegation or parallel execution is being considered.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 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.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use a single agent unless at least one condition holds:

    line 19
  2. Distinctive:

    Do not delegate a small edit, a sequential dependency chain, or work whose integration needs continuous shared context.

    line 25
  3. Distinctive:

    Never end a turn while delegated work is outstanding.

    line 39

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Delegation Discipline

One agent by default. Delegation is an optimization for genuinely independent work, not a quality badge, and never a way to bypass approvals or expand scope.

Selection gate

Use a single agent unless at least one condition holds:

  • A bounded investigation would consume substantial context and can return a concise evidence summary.
  • Two or more subtasks are independent, have non-overlapping ownership, and can proceed concurrently.
  • A specialist or independent review is required by the repository or by the change's risk.

Do not delegate a small edit, a sequential dependency chain, or work whose integration needs continuous shared context.

Cost check

Expected value must exceed setup, duplicated reading, communication, review, and merge costs. If ownership cannot be stated in one sentence, the split is not ready.

Delegation contract

Give each worker one concrete outcome; explicit file or subsystem ownership; acceptance criteria, constraints, and permitted side effects; the minimum raw context needed, without leading it to a preferred conclusion; and a required return format: findings or files changed, tests run, risks, unresolved questions.

Workers preserve other agents' and the user's changes and avoid overlapping writes. The integrating agent owns final correctness, security review, conflict resolution, and verification. Delegated results are evidence to check, not conclusions to repeat.

Completion

Never end a turn while delegated work is outstanding; collect every result, integrate, verify, then report. Return to one agent when tasks become coupled, agents duplicate investigation, merge cost grows, or the next decision needs the same context.

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.