// agents · a5c-ai/babysitter

babysitter

babysitter ist ein Open-Source-Projekt aus der Kategorie KI-Agenten von a5c-ai (a5c-ai/babysitter auf GitHub). Es hat 1.837 Sterne auf GitHub und 113 Forks, wird unter der Lizenz MIT veröffentlicht und wurde zuletzt am 16. September 2026 auf GitHub aktualisiert.

Babysitter enforces the process: deterministic, quality-gated orchestration for fleets of coding agents.

★ 1,8 Tsd. Sterne⑂ 113 ForksMITaktualisiert Sep 2026

Auf GitHub ansehen ↗Official site ↗

▦ Visual — das README dieses Projekts enthält keinen einbettbaren Screenshot,
darum hier ein Platzhalter. Die Anleitung unten ist trotzdem vollständig.

// projektdaten

Projektdaten

Repository: a5c-ai/babysitter
GitHub-Sterne: 1.837
Forks: 113
Lizenz: MIT
Zuletzt auf GitHub aktualisiert: 16. September 2026
Kategorie: KI-Agenten
Eintragstyp: Vollständige Installationsanleitung
Offizielle Website: Official site

// so funktioniert es

So funktioniert es

Clarify your goal
Decide which coding harness you will run under Babysitter and how much human oversight (breakpoints) your workflow needs.
Get the project
Install the main Babysitter CLI globally from npm, and optionally the adapters CLI and runtime platform packages.
Install and configure
Install the plugin for your harness, then run the user and project setup commands to create your profile and configure project workflows.
Run your first task
Start an orchestration run in interactive, planning, autonomous, or continuous mode and monitor it with the built-in commands.

// funktionen

Warum es so beliebt ist

Process enforcement with quality gates

Workflows defined in code are enforced step by step; progression requires gates to pass and approvals at configured breakpoints.

Immutable decision journal

Every decision is recorded, giving an auditable trail for complex, long-running agent work.

Harness-agnostic adapters

The same processes run across 12 supported AI coding harnesses, avoiding lock-in to one vendor's agent tool.

Four run modes

Interactive mode pauses for approval, planning mode stops after design, autonomous (yolo) mode runs trusted tasks fully, and continuous mode runs indefinitely for monitoring-style work.

Internal headless harness

A built-in harness runs processes programmatically without any external coding agent, suiting CI/CD and scripted orchestration, and can delegate to discovered harnesses.

Run management tooling

Doctor diagnostics, a live observation dashboard, resume for interrupted runs, retrospectives, and cleanup of old runs.

// selbst ausführen

Führe es selbst aus

Clone the project (optional)
Most users install the published npm packages instead.
```
git clone https://github.com/a5c-ai/babysitter.git
cd babysitter
```

Install the CLI and harness plugin

Install the main CLI globally, then add the plugin for your harness, for example Claude Code.

npm install -g @a5c-ai/babysitter
claude plugin marketplace add a5c-ai/babysitter-claude
claude plugin install --scope user babysitter@a5c.ai

Run your first orchestrated task
Inside your harness, invoke the babysit skill on a real task.
```
claude "/babysitter:call implement user authentication with TDD"
```

// anwendungsfälle

Was du damit bauen kannst

Running a multi-step feature build where tests must pass before the agent proceeds
Giving auditors a complete journal of what a fleet of agents decided and did
Standardising one delivery process across teams that use different coding agents
Headless CI orchestration that delegates sub-tasks to whichever harness is installed

// häufige fragen

Häufige Fragen

Who maintains it?

It is maintained by a5c-ai on GitHub.

Is it free to use?

Yes, Babysitter is MIT-licensed and free. The agents it orchestrates use your existing harness subscriptions or API keys.

Does it need a paid API key?

Babysitter itself needs no key; it drives harnesses such as Claude Code or Codex, which authenticate with your own accounts.

How does it compare to a well-known alternative?

Free-form agent chats improvise their process; Babysitter constrains agents to a coded workflow with gates, breakpoints, and a journal, trading flexibility for determinism and auditability.

How do I get started?

npm install -g @a5c-ai/babysitter, install your harness plugin, run the user and project setup commands, then start a run with /babysitter:call.

// weiter entdecken

Alternativen im Bereich KI-Agenten

Die anderen Projekte der Kategorie KI-Agenten mit den meisten Sternen im LibreAI-Katalog — die realistische Auswahl, falls babysitter nicht das Richtige ist.

openclaw★ 391,6 Tsd.
A personal AI that really does things — on any OS, with your data staying yours.
superpowers★ 294,9 Tsd.
An agentic skills framework & software development methodology that works.
ECC★ 275,2 Tsd.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
hermes-agent★ 252,0 Tsd.
The agent that grows with you — a Python framework for adaptive AI agents.
deepseek-harness★ 245,5 Tsd.
DeepSeek's plugin-based agent harness where everything is a plugin.

// mehr in agents

Mehr in KI-Agenten

Verwandte Projekte der Kategorie KI-Agenten im Katalog, die babysitter bei den Sternen am nächsten stehen — zum Weiterstöbern auf beiden Seiten.

trpc-agent-go★ 1,8 Tsd.
Go framework for production agent systems: graphs, tools, memory, A2A, MCP.
nullhub★ 1,8 Tsd.
Management console for installing and monitoring AI agents and pipelines.
gortex★ 1,9 Tsd.
Gortex, in Go — built for high-performance code-intelligence engine for ai agents.
OpenGUI★ 1,8 Tsd.
Android GUI agent framework: AI that sees, plans, and taps real mobile apps.
neo-chat★ 1,9 Tsd.
Local-first AI chat workspace for models.
CoreCoder★ 1,8 Tsd.
Self-hosted ai agents: minimal ai coding agent (~1,000 lines of python) inspir.

Hol dir babysitter

Kostenlos und offen — hol dir den Quellcode, gib dem Repository einen Stern und leg los.

Auf GitHub ansehen ↗← Mehr Tools: KI-Agenten