Solus Adopts Formal AI and LLM Contribution Policy for Open-Source Development

Solus now permits AI-assisted code and package contributions under human accountability, mandatory testing, transparent attribution, and strict rules for project communications.

Solus Formalizes Its Approach to AI-Assisted Development

On September 26, 2026, Solus published a formal AI policy describing how contributors may use artificial intelligence and large language model tools during software development. The policy frames AI systems as tools rather than authors and keeps responsibility with the human contributor.

The project says its approach is designed around transparency, accountability, quality, and maintainability. Solus also notes that it is a small volunteer team and that the policy may change as AI-assisted development evolves.

AI-Assisted Code Contributions Are Allowed

The policy does not ban AI-assisted coding. Instead, it establishes conditions for using these tools in contributions across Solus repositories, including package and tooling work.

• The contributor must test the work before submitting it.
• Non-trivial AI-assisted code must be built, installed, and tested by the contributor.
• Substantial refactors, new features, and behavioral changes require thorough testing.
• The contributor must verify that the change works as intended on Solus.

Solus deliberately leaves the definition of “trivial” somewhat flexible. The policy gives examples such as typo fixes, grammar corrections, and comment updates, while advising contributors to test when they are uncertain.

Human Accountability Remains Central

A major part of the policy is the distinction between using AI as a development aid and treating an AI system as the author. Solus states that only humans can be authors and ultimately accountable for contributions.

This means a developer cannot simply submit generated code without understanding and validating it. The human contributor remains responsible for the implementation, testing, review, and submission process.

AI Attribution: The “Assisted-by: LLM” Trailer

When AI tools assist with code that ends up in a commit, Solus requires an attribution trailer. The documented format is:

The requirement applies when AI writes or generates code that is committed, substantially modifies or refactors existing code, or produces an implementation that the contributor later adapts.

Not every interaction with an AI tool requires the trailer. Solus says research-only use, such as asking questions or exploring APIs when the human writes all of the code, does not require attribution. Mechanical tasks such as formatting or running commands are also outside the attribution requirement.

AI Bots Cannot Submit Pull Requests

Solus requires pull requests to be submitted by the developer rather than an AI bot. The contributor’s test plan should reflect that the changes were actually built, installed, and verified.

The policy also says AI planning or prompt context can be shared in pull-request comments when useful for reviewers, but AI planning documents should not be committed to the source repository.

Strict Rules for Blog Posts and Official Communications

The rules are different for project communications. Solus states that AI use for blog posts, release announcements, and other official communications is limited to translation and formatting.

AI tools must not be used to generate the content, draft narratives, or write the actual copy. The project says even partially AI-generated content is not acceptable for its blog or official communications, with the stated goal of preserving human authorship and the project’s own voice.

Transparency Is an Enforcement Requirement

Solus explicitly addresses attempts to hide AI use. Intentional obfuscation of AI tooling is grounds for rejection and may lead to a contributor being blocked from future contributions.

The stated objective is not to monitor every keystroke. Instead, the policy emphasizes honest disclosure so maintainers can evaluate contributions with appropriate context.

The Policy Is Tool-Agnostic

The rules are not tied to a particular vendor or model. Solus explicitly says the policy applies to tools such as Claude, Gemini, Copilot, ChatGPT, locally hosted models, and future systems.

For developers, this makes the important question less about which AI product was used and more about whether the contribution was reviewed, tested, attributed when required, and submitted by a human.

Why This Matters for Open-Source Maintainers

AI-assisted development can increase the speed at which code is produced, but maintainers still have to review what enters a project. Solus’s policy puts the validation burden on the person submitting the change rather than shifting it to maintainers.

This is particularly relevant for a curated operating system, where changes can affect packages, tooling, dependencies, and the daily computing environment of users. Solus’s existing packaging documentation also emphasizes controlled testing and repository processes for maintaining its rolling-release model.

What Developers Contributing to Solus Should Remember

• Use AI as a tool, not as a substitute for understanding the change.
• Build, install, and test non-trivial AI-assisted code before submission.
• Add Assisted-by: LLM when AI-generated or substantially AI-modified code enters a commit.
• Submit pull requests yourself; automated AI bots cannot submit them.
• Do not conceal AI assistance when attribution is required.
• For Solus communications, limit AI assistance to the translation and formatting uses allowed by the policy.
• Check the current policy because Solus describes it as a living document.

What’s Next for the Policy?

Solus calls the document a living policy and says it may change as AI tools and the project’s experience with them evolve. Contributors therefore need to revisit the guidance rather than treating the September 2026 version as permanently fixed.