Knowledge base (AI agents)
Everything you need to know about current AI agents and AI tools
Google Jules
Google Jules is a commercial agent tool for software development and iterative code improvement.
Description
Google Jules is a commercial agent tool for software development with a focus on terminals. It can typically understand code, suggest changes, and improve iteratively, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Zed
Zed is a commercial agent tool for software development and code review in the IDE.
Description
Zed is a commercial agent tool for software development with a focus on IDE. It can typically understand code, suggest changes, and improve iteratively, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
OpenHands
OpenHands is a platform for software development agents that autonomously handle repository tasks and automation.
Description
Platform for software development agents that can edit a repository and execute commands. The goal is to complete tasks such as setup, bug fixing, and feature work as independently as possible, including iteration via logs and tests.
area of application
Autonomous repo tasks, bug fixes, feature implementation, build and test automation, developer support.
benefits
Open source and flexible in model selection. Suitable as a basis for your own Dev Agent systems. Supports longer runs and iteration.
disadvantages
High security requirements due to code execution. Setup is challenging. Without guardrails, uncontrolled changes and costs arise.
Opendevin
Opendevin is an open source agent tool for software development with a focus on pull requests and code review.
Description
Opendevin is an open source agent tool for software development with a focus on pull requests. It can typically understand code, suggest changes, and improve iteratively, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples. The project is hosted as a repository at OpenDevin and is called OpenDevin.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/opendevin#article","headline":"Opendevin","description":"Opendevin ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/opendevin","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>Aider
Aider is a terminal tool for pair programming with Git workflow and controlled code changes.
Description
Terminal tool for pair programming with Git workflow. It accepts tasks, suggests changes across multiple files, and can prepare commits. Ideal if you work in the terminal and want to apply changes in a controlled manner using Git.
area of application
Refactoring, feature implementation, bug fixes, technical debt, rapid code changes with Git control.
benefits
Integrates very well into existing Git workflows. Makes changes traceable. Fast for iterative improvements.
disadvantages
Not fully autonomous, requires active control. Without testing, it may suggest incorrect changes. Terminal access requires secrets discipline.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/aider#article","headline":"Aider","description":"Aider ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/aider","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>Microsoft Copilot Studio
Microsoft Copilot Studio is a commercial agent tool for software development and terminal-based engineering tasks.
Description
Microsoft Copilot Studio is a commercial agent tool for software development with a focus on terminals. It can typically modify files, run tests, and summarize results, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/microsoft-copilot-studio#article","headline":"Microsoft Copilot Studio","description":"Microsoft Copilot Studio ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/microsoft-copilot-studio","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>Copilot MCP
Copilot-MCP is an open source agent tool for software development with a focus on pull requests and code review.
Description
Copilot MCP is an open source agent tool for software development with a focus on pull requests. It can typically modify files, run tests, and summarize results, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples. The project is hosted as a repository at cp mcp and is called copilot mcp.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/copilot-mcp#article","headline":"Copilot-MCP","description":"Copilot-MCP ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/copilot-mcp","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>OpenAI Codex
OpenAI Codex is a coding agent that plans tasks, writes code, and iteratively improves results for engineering workflows.
Description
Coding agent that plans tasks, writes code, executes commands, and iteratively improves results. The goal is to perform real engineering steps, not just deliver code snippets. Suitable for repo work when tests and acceptance criteria exist.
area of application
Feature implementation, refactoring, test additions, debugging, code review preparation, rapid prototyping.
benefits
High leverage for routine coding tasks. Works step by step and can take tests into account. Good basis for repeatable engineering workflows.
disadvantages
Errors remain possible, so review is mandatory. Access to repo and secrets must be secured. Costs and limits may be relevant.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/openai-codex#article","headline":"OpenAI Codex","description":"OpenAI Codex ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/openai-codex","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>SWE agent
SWE-agent is an open source agent tool for software development with a focus on pull requests and code review.
Description
SWE agent is an open source agent tool for software development with a focus on pull requests. It can typically understand code, suggest changes, and improve iteratively, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples. The project is hosted as a repository at SWE agent and is called SWE agent.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/swe-agent#article","headline":"SWE-agent","description":"SWE-agent ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/swe-agent","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>Devin
Devin is a commercial software engineer agent that autonomously handles tickets, code, and pull requests.
Description
Commercial software engineer agent who tracks tickets over a longer period of time and creates code, tests, and PRs. The typical process is to create a plan, implement, test, iterate, and then hand over via pull request. Works best when tasks are clearly defined.
area of application
Autonomous processing of small to medium-sized tickets, maintenance, proof of concepts, test expansion, technical debt.
benefits
End-to-end workflow saves time in everyday life. Can process multiple steps without constant prompting. Useful for reducing backlogs.
disadvantages
Costs and vendor dependency. Data protection and IP issues must be clarified. Risky without review and branch protection.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/devin#article","headline":"Devin","description":"Devin ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/devin","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>TheiaAI TheiaIDE
TheiaAI TheiaIDE is a commercial agent tool for software development and code review in the IDE.
Description
TheiaAI TheiaIDE is a commercial agent tool for software development with a focus on IDE. It can typically understand code, suggest changes, and improve iteratively, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/theiaai-theiaide#article","headline":"TheiaAI TheiaIDE","description":"TheiaAI TheiaIDE ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/theiaai-theiaide","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>Amazon Q CLI
Amazon Q CLI is an open-source agent tool for software development and pull request management.
Description
Amazon Q CLI is an open source agent tool for software development with a focus on pull requests. It can typically modify files, run tests, and summarize results, working within the context of your repository. The goal is to process engineering steps rather than just providing code examples. The project is hosted as a repository at aws and is called amazon q developer cli.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/amazon-q-cli#article","headline":"Amazon Q CLI","description":"Amazon Q CLI ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/amazon-q-cli","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>Gemini CLI
Gemini CLI is an open source agent tool for software development and pull request management.
Description
Gemini CLI is an open source agent tool for software development with a focus on pull requests. It can typically understand code, suggest changes, and improve iteratively, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples. The project is hosted as a repository at google gemini and is called gemini cli.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/gemini-cli#article","headline":"Gemini CLI","description":"Gemini CLI ist ein Open Source Baustein für Agent Automatisierung und Infrastruktur mit Schwerpunkt auf Framework. Es stellt entweder eine Laufzeit, ein Workflow System, eine Integrationsschicht oder Messbarkeit bereit, damit Agents zuverlässig mit Tools und Daten arbeiten.","url":"https://www.lj-partner.com/ki-agents/gemini-cli","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Automatisierung"}</script>Cline
Cline is a VS Code agent that can modify files and execute shell commands for iterative development.
Description
VS Code Agent, which can modify files and execute shell commands. Supports tool integration, often via MCP, and works directly in the project directory. Good for iterative development if you manage the agent closely and review changes.
area of application
Coding in the editor, refactoring, debugging, testing, setting up projects, quickly implementing small tasks.
benefits
Highly productive in daily IDE workflow. Can use context from the repository. Tool integration enables structured actions.
disadvantages
Without a sandbox and permissions, the concept is very risky. May unintentionally modify files. Requires a clear review process and restrictions.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/cline#article","headline":"Cline","description":"VS Code Agent, der Dateien ändern und Shell Befehle ausführen kann. Unterstützt Tool Integration, oft über MCP, und arbeitet direkt im Projektverzeichnis.","url":"https://www.lj-partner.com/ki-agents/cline","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Softwareentwicklung"}</script>Cursor Agent
Cursor Agent is a commercial tool for automating software development tasks in your IDE.
Description
Cursor Agent is a commercial agent tool for software development with a focus on IDE. It can typically modify files, start tests, and summarize results, working within the context of your repository. The goal is to process engineering steps rather than just providing code examples.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/cursor-agent#article","headline":"Cursor Agent","description":"Cursor Agent ist ein Kommerziell Agent Werkzeug für Softwareentwicklung mit Fokus auf IDE. Es kann typischerweise Dateien ändern, Tests starten und Ergebnisse zusammenfassen und arbeitet dabei im Kontext deines Repos.","url":"https://www.lj-partner.com/ki-agents/cursor-agent","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Softwareentwicklung"}</script>Emacs MCP
Emacs MCP is an open-source agent tool for code review and iterative improvements in pull requests.
Description
Emacs MCP is an open source agent tool for software development with a focus on pull requests. It can typically understand code, suggest changes, and improve iteratively, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples. The project is located in the emacs mcp repository and is called emacs mcp.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/emacs-mcp#article","headline":"Emacs MCP","description":"Emacs MCP ist ein Open Source Agent Werkzeug für Softwareentwicklung mit Fokus auf Pull Requests. Es kann typischerweise Code verstehen, Änderungen vorschlagen und iterativ verbessern und arbeitet dabei im Kontext deines Repos.","url":"https://www.lj-partner.com/ki-agents/emacs-mcp","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Softwareentwicklung"}</script>Goose
Goose is an open source agent tool for software development with a focus on pull requests and code changes.
Description
Goose is an open source agent tool for software development with a focus on pull requests. It can typically modify files, run tests, and summarize results, working within the context of your repository. The goal is to process engineering steps rather than just providing code examples. The project is stored as a repository at block and is called goose.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/goose#article","headline":"Goose","description":"Goose ist ein Open Source Agent Werkzeug für Softwareentwicklung mit Fokus auf Pull Requests. Es kann typischerweise Dateien ändern, Tests starten und Ergebnisse zusammenfassen und arbeitet dabei im Kontext deines Repos.","url":"https://www.lj-partner.com/ki-agents/goose","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Softwareentwicklung"}</script>VS Code GitHub Copilot
VS Code GitHub Copilot is a commercial agent tool for automating software development tasks in your repository.
Description
VS Code GitHub Copilot is a commercial agent tool for software development with a focus on the terminal. It can typically modify files, run tests, and summarize results, working within the context of your repository. The goal is to work through engineering steps rather than just providing code examples.
area of application
Bug fixes, refactoring, test additions, CI debugging, minor features, PR preparation, and code review support.
benefits
Saves time on routine work, especially when tests are available. Good control when changes go through Git Review. Increases throughput for tickets with clear acceptance criteria.
disadvantages
Without testing and reviews, the risk of incorrect changes increases. Access to repositories and secrets must be strictly limited. Model limits and costs can slow down engineering flows.
Schema (copy & paste)
<script type="application/ld+json">{"@context":"https://schema.org","@type":"Article","@id":"https://www.lj-partner.com/ki-agents/vs-code-github-copilot#article","headline":"VS Code GitHub Copilot","description":"VS Code GitHub Copilot ist ein Kommerziell Agent Werkzeug für Softwareentwicklung mit Fokus auf Terminal. Es kann typischerweise Dateien ändern, Tests starten und Ergebnisse zusammenfassen und arbeitet dabei im Kontext deines Repos.","url":"https://www.lj-partner.com/ki-agents/vs-code-github-copilot","inLanguage":"de","author":{"@type":"Person","name":"Lars Jurischka"},"publisher":{"@type":"Organization","name":"LJ & Partner","url":"https://www.lj-partner.com"},"articleSection":"Softwareentwicklung"}</script>
Book now