agentleFS
D-SheyD-Shey/autoharness/.agents/skills/multiagent-orchestration/SKILL.md

multiagent-orchestration

Skill10 starsChanged todayMITView on GitHub

Role-based multi-agent coordination with isolated subagents and gated milestone checks

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:

    Use when dividing work among multiple specialized agent roles (e.g.

    line 9
  2. Distinctive:

    Run subagents with independent context windows to prevent token exhaustion and distraction.

    line 18
  3. Very specific:

    Never mark a subagent task complete without automated verification pass (autoharness verify).

    line 23
  4. Distinctive:

    Run a multi-step task involving 2+ distinct roles.

    line 28
  5. Very specific:

    Verify all role actions and artifacts are logged in .autoharness/swarm/ and searchable in memory.

    line 31

≈ “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

multiagent-orchestration

When to use

Use when dividing work among multiple specialized agent roles (e.g. Architect, Researcher, Worker/Coder, Verifier, Critic) to prevent single-context bloat and enforce independent review gates.

Steps

  1. Define specialized roles for subtasks:
    • researcher: reads codebase and external docs, generates specification or evidence.
    • coder: implements modifications according to specification.
    • verifier: executes tests and checks invariants.
    • critic: independent evaluation without author bias.
  2. Isolate agent workspace context:
    • Run subagents with independent context windows to prevent token exhaustion and distraction.
    • Communicate between agents strictly via structured inputs/outputs (JSON artifacts or memory records).
  3. Connect subagents to shared Gateway memory:
    • Index intermediate results in .autoharness/memory.db so subsequent agents can retrieve them using BM25 query.
  4. Enforce Hard Verification Gates:
    • Never mark a subagent task complete without automated verification pass (autoharness verify).
    • Escalate to arbitrator if a subtask fails verification multiple times.
  5. Aggregate final results into an executive summary.

How to verify

  1. Run a multi-step task involving 2+ distinct roles.
  2. Confirm each role receives only necessary inputs, not the entire conversation history.
  3. Confirm that verification gate prevents invalid intermediate state from propagating to downstream workers.
  4. Verify all role actions and artifacts are logged in .autoharness/swarm/ and searchable in memory.

Facts found here

  • Role specialization prevents context degradation in multi-step workflows (as demonstrated by ByteDance Deer-Flow and Ruflo).

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.