koog
Koog is a JVM (Java and Kotlin) framework for building predictable, fault-tolerant and enterprise-ready AI agents across all platforms – from backend services to Android and iOS, JVM, and even in-browser environments. Koog is based on our AI products expertise and provides proven solutions for complex LLM and AI problems
No outside reports on file
Koog is a framework tool, so security depends heavily on how developers use it and what permissions they grant to AI agents running within it.
Based on public adoption and transparency signals as of Oct 21, 2026. Our opinion, not a code audit or security guarantee.
What it can access
- Local filesNo local folder access notedLOW
- Command executionNo shell command runner notedLOW
- Network reachFetches arbitrary web URLsREVIEW
- CredentialsNo API token requiredLOW
- Package sourceGitHub (JetBrains/koog)PUBLIC
Inferred from this listing’s public package details and description. Check the project’s source before installing.
Install
No npm or PyPI package is listed. Install from source: github.com/JetBrains/koog ↗
Everything known
We found nothing about this project in the sources we read (last checked 2026-10-30). That does not mean it has no problems.
0 incidents · 0 advisories · 0 news articles · 10 community threads · checked 2026-10-30
Popularity grade: C (from public signals such as stars and downloads, not a code review). Graded 2026-10-21.
Showing all 10Showing 10 community threads of 10
- CommunityDoes JetBrains Koog support Claude Agent Skills (skills.md)?
- CommunityDoes koog support iOS targets in Kotlin Multiplatform projects?
- CommunityDoes koog support local LLMs on Android or iOS?
- CommunityDoes Koog support the A2A Protocol for multi-agent systems?
- CommunityHow to call Gemini via Google Vertex AI in JetBrains Koog without a Vertex executor?
- CommunityHow to use LiteRT / on-device Gemma 3n models in JetBrains Koog?
- Communitykoog + LM Studio: listing models fails with "Field 'created' is required for OpenAIModel"
- CommunityKoog crashes with NoSuchMethodError after kotlinx-datetime 0.7.1
- Communitykoog executeStreaming() missing tools arg: tool calls dropped in streaming responses
- CommunityKoog: can I use tool calls with executeStreaming?
Sources we read: OSV, GitHub Advisory Database, CISA Known Exploited Vulnerabilities, 44 security news feeds and our incident library. Last checked 2026-10-30.
About Koog
Koog is a framework for building AI agents using Java and Kotlin programming languages. It lets developers create agents that can run reliably on different platforms (servers, Android, iOS, web browsers) and handles common problems with large language models.
Who it’s for
Java and Kotlin developers who want to build AI agents for enterprise applications. Teams that need AI systems to work predictably across multiple platforms and environments.
Alternatives
- LangChain: Multi-language framework for building AI agents with many integrations
- Semantic Kernel: Microsoft's framework for AI agents (supports multiple languages)
- AutoGen: Framework for multi-agent AI systems with conversation patterns
- Verba: Open-source framework focused on retrieval-augmented generation for agents
Setup
Check these steps against the project’s own instructions before running anything.
Follow the project's README at https://github.com/JetBrains/koog (opens external site) for installation and configuration instructions specific to your platform and use case.
Why this grade
Signal-based grade (3/7 points): well starred on GitHub, public source code. This reflects adoption and transparency, not a code review.
Popularity and upkeep
- GitHub stars
- 4.6k
- Listing last refreshed
- 2026-10-30
These are public popularity signals, not a maintenance check. Look at the repository’s commit history and open issues.
Risks to watch for
- LLM prompt injection: AI agents can be tricked by malicious text in user inputs to perform unintended actions or leak information
- Platform access: Agents could potentially access files, databases, or services available to the JVM process they run in
- Model behavior uncertainty: Language models sometimes produce unexpected outputs that the framework may not catch, leading to wrong decisions
- Dependency vulnerabilities: Koog relies on other Java libraries that could have security flaws affecting all applications using it
- Cross-platform exposure: Running on Android, iOS, and web means security issues could affect multiple platforms through shared code
How to lower the risk
Least-privilege permissions
- Grant AI agents only the minimum access they need to function
- Avoid giving agents direct database or file system access when possible
Network isolation
- Run agents in sandboxed environments if handling untrusted input
- Restrict which external services agents can connect to
Monitoring
- Log all agent decisions and actions for review
- Alert on unusual patterns or unexpected agent behavior
Red flags to watch for
- Agents returning data they shouldn't have access to
- Unexpected or evasive responses to simple requests
- Performance issues suggesting resource abuse
Questions
How do I install Koog?
Koog does not publish a standalone npm or PyPI package. Clone or follow the setup instructions in its source repository at github.com/JetBrains/koog.
Are there security reports about Koog?
No reports found for this package in our research (last checked 2026-10-30). Koog currently holds Grade C (Moderate risk signals) based on public adoption and transparency signals. Key risk areas to watch include llm prompt injection, platform access, model behavior uncertainty.
Is Koog actively maintained?
Koog has 4.6k GitHub stars. Our directory data was last refreshed on 2026-10-30. Check recent commit activity, release tags, and open issues at github.com/JetBrains/koog before pinning it in production.
What are alternatives to Koog?
Notable alternatives to Koog include LangChain (Multi-language framework for building AI agents with many integrations), Semantic Kernel (Microsoft's framework for AI agents (supports multiple languages)), AutoGen (Framework for multi-agent AI systems with conversation patterns), Verba (Open-source framework focused on retrieval-augmented generation for agents). You can also compare Koog side by side with other MCP servers in the Related tools section on this page.
Key facts
- Type
- App connector
- Publisher
- JetBrains
- Source code
- github.com/JetBrains/koog
- Est. context load
- ~750 tokens / turn (~4 tool schemas, 0.38% of 200k context)
- Popularity
- 4.6k GitHub stars
- Last updated
- 2026-10-30
- Last graded
- 2026-10-21
Community
Discussions for Koog
Ask questions or share configuration tips. Governed by our Community Rules & Terms.
No posts match this filter yet.
Be the first to ask a question or share a working tip.
Check before you install, in Claude
Your assistant answers Block, Review or No known issues for any package before it installs it.
Get this check inside Claude