Recipes / Productivity

Our rating: 2/5 for this setup

A Manus-style agent team built from Claude Code skills

A newsletter writer stopped paying for Manus credits and built a Manus-like setup inside Claude Code from 10 skills that act as a team. An orchestrator hands each request to a planner, and long jobs run in the background with checkpoints every 30 minutes. A code generator builds models in Python, and a quality checker scores the result before it reaches you.

New Added Nov 2, 2026 · ✓ Last checked Nov 2, 2026

It broke for meBroke?
How it works: A Manus-style agent team built from Claude Code skills. Trigger: You give Claude Code a multi-step task. The bot (Claude Code with 10 custom skills (orchestrator, planner, quality checker and more)): The orchestrator and planner split it into stages; Background agents research, write code and draft; A quality checker scores the result. Result: A finished report and a notification. Replaces Manus AI.HOW IT WORKSA Manus-style agent team built from Claude Code skills1 · TRIGGER2 · THE BOT3 · RESULTYou give ClaudeCode a multi-steptaskClaude Code with 10 cus…The orchestrator and planner splitit into stagesBackground agents research, writecode and draftA quality checker scores the resultA finished reportand a notificationREPLACESManus AIRATED 2/5 · HARD SETUPAppToBot
How it works: trigger: You give Claude Code a multi-step task; the bot (Claude Code with 10 custom skills (orchestrator, planner, quality checker and more)): The orchestrator and planner split it into stages; Background agents research, write code and draft; A quality checker scores the result; result: A finished report and a notification.HOW IT WORKS1 · TRIGGERYou give Claude Code amulti-step task2 · THE BOTClaude Code with 10 cus…The orchestrator and planner split itinto stagesBackground agents research, write codeand draftA quality checker scores the result3 · RESULTA finished report and anotificationREPLACESManus AI
Diagram drawn by AppToBot from this recipe's source. It shows the flow, not a screenshot of the product.

Setup steps

  1. Create an orchestrator skill that receives requests and delegates them
  2. Add planner, workflow designer, progress tracker, quality checker and pattern learner skills
  3. Add an async executor, code generator, long task manager and event stream for heavy jobs
  4. Give it a multi-step task and let long jobs run in the background
  5. Review the report when the notification arrives

Steps summarized from the original Blog source ↗. Use “Copy” to grab them with the tools and source link, or copy an AI prompt.

How it went

The writer's first test, a market research task with a financial model and an investment report, came back in 42 minutes without them touching it, and they rated it better than anything their old setup produced.

Worth knowing

The 92% quality score comes from the system's own checker, not an outside review. The roughly 1,000 lines of prompts are only for subscribers, and the post links to Manus with a referral code.

Only what the original source ↗ says, re-read Nov 2, 2026. How we write these

Try it with AI

Paste this into ChatGPT, Claude or your own agent for step-by-step help. Prompt written by AppToBot from this recipe's data, not by the original author. It keeps drafts for your review and asks before anything is sent or paid.

Help me set up this automation using Claude Code with 10 custom skills (orchestrator, planner, quality checker and more): "A Manus-style agent team built from Claude Code skills".
Goal: replace Manus AI.

How it should work:
- Trigger: You give Claude Code a multi-step task
- Then: The orchestrator and planner split it into stages
- Then: Background agents research, write code and draft
- Then: A quality checker scores the result
- Result: A finished report and a notification

Setup steps (summarized from the original source):
1. Create an orchestrator skill that receives requests and delegates them
2. Add planner, workflow designer, progress tracker, quality checker and pattern learner skills
3. Add an async executor, code generator, long task manager and event stream for heavy jobs
4. Give it a multi-step task and let long jobs run in the background
5. Review the report when the notification arrives

Guide me one step at a time. Ask me for anything you need (accounts, sheet or channel names) before you start.

Safety rules:
- Build and test with sample data first, and show me a sample result before anything goes live.
- Keep anything that goes to other people (emails, posts, replies, record updates) as drafts for my review until I say otherwise.
- Never send, post, buy, pay, delete or change records without my approval.
- Tell me about any cost (paid plans, API usage) before I'd incur it, and help me set a spending limit.
- Keep API keys and passwords out of shared docs and chats.
- Privacy: Runs in your Claude Code session, so task data goes to Anthropic.

Based on: https://www.restless-brain.com/p/how-i-stopped-paying-for-manus-ai

Manus AIClaude Code with 10 custom skills (orchestrator, planner, quality checker and more)

Tool names link to each official site (opens in a new tab).

Cost afterNot reported
SetupHard · Not reported
Time savedNot reported
SkillAdvanced

Running cost: Not reported The writer says it runs 'for free' in Claude Code but doesn't say which Claude plan they use. How we sort costs
Stack: Claude

App vs bot, side by side

Old app vs bot
Old app / habitBot
ToolManusClaude Code skills that plan, research, code, write and check quality
Monthly costCredit-based; hard tasks use many credits (per writer, no amount)Not reported
SetupNone (already in use)Not reported · Advanced
Your timeBaselineNot reported
MaintenanceHandled by vendor/you as beforeYou maintain about 1,000 lines of prompts and instructions.
PrivacyData stays with the app vendorRuns in your Claude Code session, so task data goes to Anthropic.

Full comparison

ReplacesManus
Bot / agent stackClaude Code skills that plan, research, code, write and check quality
Monthly cost beforeCredit-based; hard tasks use many credits (per writer, no amount)
Monthly cost afterNot reported
Setup timeNot reported
Time savedNot reported
Skill levelAdvanced
Reliability2/5: one test, graded by the system's own quality checker, and the setup instructions are gated.
MaintenanceYou maintain about 1,000 lines of prompts and instructions.
Privacy & dataRuns in your Claude Code session, so task data goes to Anthropic.
Best forClaude Code users who keep running long multi-step research jobs
Not forPeople who want a hosted agent that just works

Pros and cons

Pros

  • No per-task credits
  • Long jobs keep running in the background with checkpoints
  • You can read and change every agent's instructions

Cons

  • One test, scored by its own checker
  • Full setup instructions only for subscribers

Source and evidence

Newsletter post (Nov 2025) with one test: a market research, financial model and report task finished in 42 minutes with a 92% score from the system's own checker. The full instructions are for subscribers, and the post includes a Manus referral link.

Read the original Blog source ↗

“Not reported” means the source doesn't say. We never fill gaps with guesses. Numbers marked “per author” or “vendor claim” are the source's own and weren't independently tested. Source last checked Nov 2, 2026.

Our rating is an editorial opinion based on the public source above, not lab testing. Spot an error? Request a correction.

Plain-text summary (for search and AI assistants)

A Manus-style agent team built from Claude Code skills. Replaces Manus with Claude Code skills that plan, research, code, write and check quality. Cost before: Credit-based; hard tasks use many credits (per writer, no amount). Cost after: Not reported. Setup: Not reported, Advanced. Time saved: Not reported. Reliability 2/5: 2/5: one test, graded by the system's own quality checker, and the setup instructions are gated. Best for: Claude Code users who keep running long multi-step research jobs. Not for: People who want a hosted agent that just works. How it went (per source): The writer's first test, a market research task with a financial model and an investment report, came back in 42 minutes without them touching it, and they rated it better than anything their old setup produced. Worth knowing: The 92% quality score comes from the system's own checker, not an outside review. The roughly 1,000 lines of prompts are only for subscribers, and the post links to Manus with a referral code. Source: https://www.restless-brain.com/p/how-i-stopped-paying-for-manus-ai

Related recipes

More: all Productivity recipes

Compare with another recipe ← All recipes

Questions, corrections or ideas?

Spotted an error, want a recipe covered, or have a partnership idea? Send us a message and we'll get back to you.

Contact us

Starting your own newsletter? Try beehiiv ↗

What is beehiiv?

beehiiv is a newsletter platform: you write and send emails, host a newsletter website, and grow and monetize a list from one place. We've used it for our own newsletter.

beehiiv says (vendor claims, not tested by us):

  • Free plan with up to 2,500 subscribers and unlimited sends
  • Built-in website, recommendation network and analytics
  • 0% cut of paid subscription revenue (Stripe fees apply)

Good for: starting or growing an email newsletter. Not ideal for: email automations or removing beehiiv branding on a $0 budget (those need Lite or Pro).

Free plan: $0, no card needed. Paid plans from $49/mo billed annually (Lite, up to 1,000 subscribers). People who sign up through our link get a 14-day trial plus 20% off their first 3 months, per beehiiv's partner page. Facts as of Oct 2026, from beehiiv's own website. More about beehiiv →

Press / anywhere to search. Open full search

Send us a bot that replaced an app

Three quick fields. We check every link by hand and only publish what the source supports.

Please don't enter an email address here; we never publish emails. Stored for up to 12 months. See Privacy and Terms.