Developer Principles
This page is not an integration manual. It is AgentOS's long-term set of principles for developers.
We want Agent Apps to be not only functional, but also responsible in open source usage, compliance, ethics, and public value.
The developer ecosystem we want to encourage
We welcome developers who:
- understand their real product scenario
- use AI or agents to solve meaningful problems
- respect the open source work that came before them
- are willing to give back code, lessons, or practices when possible
- want to use AI for a more productive and hopeful future
Forks, open source licenses, and derivative work
If your app is built by forking an open source project and adding agent capabilities on top, you should follow the original open source license.
At a minimum, that means:
- keeping required copyright, license, and attribution notices
- following the original license terms for distribution, modification, and reuse
- not treating open-source-licensed code as if you could arbitrarily close or relicense it
We strongly encourage you to open source your own work as well.
But when doing so, you should still respect the original project's license boundaries and attribution.
Source links and version submissions
If your app evolves from an existing open source project, or if you are willing to share your implementation, we recommend that each new app version submission include:
- the source repository URL for the submitted version
- the original upstream project URL
- the relevant license information
- a brief explanation of the key changes you made
This helps:
- users and reviewers understand the project's origin
- reduce license ambiguity and attribution disputes
- build a healthier ecosystem of reuse and collaboration
Legal compliance and basic ethics
We want AI and agents to help create a future of much higher productivity, more light, and more hope.
Because of that, developers building Agent Apps should comply with applicable laws and regulations and stay within basic human ethical boundaries.
At a minimum, that includes:
- not creating or distributing unlawful content
- not encouraging fraud, harassment, manipulation, or harm
- not designing agents primarily for malicious or harmful uses
- not hiding behind “technology neutrality” to avoid responsibility
What we explicitly reject
We do not welcome apps that:
- primarily spread harmful or abusive information
- are built for malicious exploitation, gray-market abuse, fraud, or manipulation
- clearly violate legal or ethical boundaries
- use agent capabilities to amplify harmful behavior
Put simply, we say no to harmful content and apps driven by dark intent.
Why these principles matter
Agent Apps often have more real-world impact than a normal feature page.
That is because they may:
- call models continuously
- operate tools or external systems automatically
- execute multi-step tasks on behalf of users
- influence real outcomes across longer chains of action
That is exactly why developers should think about open source obligations, legal compliance, and ethical boundaries early.
Suggested self-check before release
- Does my app use third-party open source code, and am I following the original license?
- If I forked a project, have I clearly documented the origin and my changes?
- If I am willing to open source my work, have I prepared a repository URL to include with version submission?
- Could my app obviously be used for unlawful, fraudulent, harassing, manipulative, or otherwise harmful purposes?
- If misuse is likely, have I designed clear boundaries, restrictions, or refusals?
Next steps
If you have already decided your scenario is worth turning into an Agent App and you are comfortable following these principles, continue with: