Hermes Swarm Roles Explained: Builder, Reviewer, More

A Hermes Swarm comes with 6 preset agent roles out of the box — and picking the right ones is what makes your swarm actually work.

This post covers each of the 6 preset roles.

What each one does.

When to use it.

How to combine them for different workflows.

The 6 Preset Roles

When you click "Add swarm" in Hermes Workspace, you can pick from:

Builder
Reviewer
Triage
Lab
Sage
Scribeex

Each comes with preset system prompts and skills.

You can use them out of the box or customise.

Role 1 — Builder

The doer.

What Builder does:

Drafts content.
Writes code.
Creates files.
Produces output.

When to use Builder:

Content creation workflows.
Code generation.
Anything where you need raw output.

Best paired with: Reviewer.

Role 2 — Reviewer

The QA agent.

What Reviewer does:

Reads Builder's output.
Catches errors.
Suggests improvements.
Validates quality.

When to use Reviewer:

Always after a Builder.
Quality-critical workflows.
Production-ready content.

The Reviewer pattern is what separates good swarms from broken ones.

I cover the broader critic-step principle (same as Google Simula's dual critic filter — see Google Simula Mechanism Design) — applied here in Hermes.

Role 3 — Triage

The router.

What Triage does:

Reads incoming requests.
Decides which agent should handle.
Routes work to the right specialist.

When to use Triage:

Customer-facing workflows.
Multi-channel inputs.
When tasks vary widely.

Best paired with: every other role (acts as the front door).

Role 4 — Lab

The experimenter.

What Lab does:

Tries new approaches.
Tests hypotheses.
Generates options for evaluation.

When to use Lab:

Innovation work.
A/B testing ideas.
Early-stage exploration.

Best paired with: Sage (for evaluation).

🔥 Want my full Hermes Swarm role playbook? Inside the AI Profit Boardroom, I share my exact Hermes Swarm role configs for content, customer ops, research, and more. Plus 2-hour Hermes course and weekly live coaching. 2,800+ members. → Get the playbook

Role 5 — Sage

The advisor.

What Sage does:

Provides strategic input.
Evaluates options.
Offers wisdom on decisions.

When to use Sage:

Strategic planning.
Decision support.
High-stakes evaluation.

Best paired with: Lab (for ideating then evaluating).

Role 6 — Scribeex

The documentarian.

What Scribeex does:

Captures what happened.
Writes summaries.
Maintains records.

When to use Scribeex:

Long workflows where state matters.
Compliance-sensitive work.
When you need clear audit trails.

Best paired with: any combination of others (acts as the historian).

Common Role Combinations

For specific workflows.

Content creation swarm

Builder — drafts the content.
Reviewer — checks quality.
Scribeex — keeps records of decisions.

Customer ops swarm

Triage — routes incoming messages.
Builder — drafts responses.
Reviewer — checks before sending.

Research swarm

Lab — explores hypotheses.
Sage — evaluates findings.
Scribeex — documents conclusions.

Engineering swarm

Builder — writes code.
Reviewer — reviews PRs.
Sage — architectural decisions.

For each workflow, the right combination matters.

Custom Roles

Beyond the 6 presets, you can create custom roles.

When to go custom:

Domain-specific work (legal, medical, financial).
Unique workflow steps.
When presets don't quite fit.

How:

Click "Add swarm" → "Custom".
Define system prompt.
Set skills.
Save.

Custom roles let you build specialist agents for your specific work.

Aurora: The Orchestrator

All 6 roles report to Aurora.

Aurora is the master orchestrator:

Receives your mission.
Routes work to sub-agents.
Coordinates handoffs.
Reports progress.

You don't need to micromanage.

You set the goal. Aurora orchestrates.

How To Pick Which Roles For Your First Swarm

If you're new:

Start with Builder + Reviewer for any content/output work.
Add Triage if you have varied incoming requests.
Add Sage if you want strategic input.
Add Lab for experimentation.
Add Scribeex for documentation needs.

Don't add all 6 on day one.

Start lean. Expand as needed.

Time Savings Per Role

Approximate from my own use:

Builder swarm (alone): saves ~50% time vs solo writing.
Builder + Reviewer: saves ~70% time and improves quality.
Triage + Builder + Reviewer: saves ~80% time on customer ops.

Each role added (when used right) compounds the savings.

Common Role Mistakes

1. Skipping Reviewer.

Builder alone produces inconsistent quality.

Always add a Reviewer.

2. Adding too many roles.

5+ roles for simple tasks is overkill.

Start with 2-3.

3. Vague role definitions.

If two roles overlap, your master gets confused.

Make each role narrow and clear.

4. Forgetting handoff rules.

Tell each role how to hand off to the next.

Without explicit handoffs, agents drift.

Where Roles Limit You

Be honest.

Doesn't replace specialist domain experts.
Doesn't handle creative judgment well.
Doesn't fully replace senior decision-makers.

For execution work, roles are great.

For strategic judgment, humans still matter.

Role Setup Time

Setting up a single role: 5-10 minutes.

Setting up a 3-role swarm: 20-30 minutes.

Setting up a 6-role swarm: 60-90 minutes.

Time is upfront. Then you reuse.

How Roles Pair With Channels

You can connect Hermes Swarm to channels (Telegram, Discord, etc).

Common pattern:

Triage at the channel front door.
Builder drafts responses.
Reviewer validates before sending.

This is the same multi-agent pattern as Telegram AI Agent Architecture but applied in Hermes.

🚀 Want my full role playbook? The AI Profit Boardroom has my Hermes Swarm role templates, OpenClaw 6-hour course, 2-hour Hermes course, daily training, weekly live coaching. 2,800+ members. → Join here

FAQ — Hermes Swarm Roles

Which role should I add first?

Builder — most universal.

Do I need Reviewer?

For quality-critical work, yes.

For exploration, optional.

Can custom roles replace presets?

Yes — for specialist domains, custom often fits better.

How does Aurora know which role to call?

Aurora reads role descriptions and your mission, then routes intelligently.

Can two agents have the same role?

Yes — useful for parallelism (e.g. multiple Builders working different parts).

What if a role gets stuck?

Click the agent, check API/model settings.

Usually wrong model assigned.

Can roles share memory?

Limited by Hermes Workspace settings.

Configure per agent.