Hermes LoopHermes Loop
Library

Crew templatesAgent teams you can run, edit, or build.

Create custom crew →
What this is
A crew is an ordered sequence of specialist agents. Built-in crews ship with prompts and Zod-validated output schemas; custom crews let you compose your own.
What to do
Pick a crew that matches the work, then click Launch to start a mission.
What happens next
The mission page shows live progress, deliverables, approvals, and a workflow receipt at the end.
Markets6 agents
Paper Trading Desk

Research markets and simulate trading decisions. Simulation only — no broker, no real money.

01
Market Scout
Scans the watchlist and produces a market brief.
02
News Agent
Summarizes relevant news and event flow.
03
Strategy Agent
Builds a single, defensible trade thesis.
04
Risk Agent
Calculates downside, invalidation, and sizing notes.
05
Backtest Agent
Runs a simple historical simulation from seeded data.
06
Paper Execution Agent
Records the simulated trade decision only.
Example objective
Analyze AAPL and NVDA for a paper trade this week given the macro backdrop.
Quality6 agents
Bug Hunter Crew

Test sites like a human QA team. Demo-snapshot mode if browser automation is unavailable.

01
Explorer Agent
Maps pages and key flows via real browser audit.
02
Flow Tester Agent
Tests signup/contact/checkout-style flows.
03
Accessibility Agent
Checks labels, contrast notes, keyboard risks.
04
Copy Agent
Flags confusing or weak copy.
05
Bug Reporter Agent
Creates recommended fixes for each issue.
06
Report Agent
Drafts a client-ready QA report summary.
Example objective
Audit https://example-shop.test for conversion and accessibility issues.
Personal6 agents
Life Admin Crew

Handle annoying personal admin with human approval. Drafts only — never auto-sent.

01
Intake Agent
Identifies the task and required outcome.
02
Evidence Agent
Extracts dates, amounts, names, order numbers from docs.
03
Policy Agent
Creates a rights/policy angle.
04
Draft Agent
Writes the email/message/script.
05
Critic Agent
Checks missing info and weak points.
06
Follow-up Agent
Creates reminders and next steps.
Example objective
Help me dispute a broken-product refund from order #A4827.
Engineering5 agents
Codebase Debugger Crew

Diagnose build / typecheck / lint failures in the project workspace using the governed runtime tools. Read-only diagnosis — no file edits.

01
Repo Scout
Maps the workspace and gathers cheap diagnostic context.
02
Build Runner
Runs typecheck / lint / build and captures the failure surface.
03
Error Analyst
Groups errors by root cause and identifies the load-bearing files.
04
Fix Planner
Drafts a concrete fix plan per root cause. Read-only — no edits.
05
Report Agent
Wraps findings + plan into a deliverable.
Example objective
Run typecheck and lint diagnostics on the workspace, then report the top errors with concrete fix recommendations.
sexsexsexs4 agents
sexsexsexs

sexsexsexs

01
Research Agent
Finds the important context and constraints.
02
Strategy Agent
Turns research into a practical plan.
03
Critic Agent
Challenges weak assumptions and missing evidence.
04
Report Agent
Writes the final operator-facing summary.
Example objective
sexsexsexs
Research2 agents
Quick Audit

A general-purpose two-agent crew. The Researcher gathers context, surfaces facts, and flags gaps. The Reporter synthesizes those findings into a Markdown report with severity-tagged findings, recommended actions sorted by impact, and a confidence rating. Use it for pre-launch checks, quick competitive reviews, deal due diligence, or any "look at X and tell me what's important" ask.

01
Researcher
Gathers context and surfaces the relevant facts.
02
Reporter
Synthesizes the findings into a clean operator-facing report.
Example objective
Audit https://example.com — identify the top 3 issues, the top 3 strengths, and produce a Markdown report with prioritized recommendations.