docs: Create AI Policy

[DOT-EXECUTE]

Description

Created AGENTS.md, to clarify acceptable and unacceptable usage of AI Agents and tools. Updated README.md, to reference AGENTS.md with a 'for agents' bolded section. Closes #173

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update

Key Changes

Validation

N/A (Docs change)

Checklist

• I reviewed the diff and kept it focused on the described change.
• Python changes pass uv run ruff check . and uv run ruff format --check ., or this PR does not change Python.
• Changed behavior is covered by appropriate tests, including a regression test for bug fixes; any limitations are explained under Validation.
• Validation lists the checks or manual flows I ran and their results.
• Relevant documentation is updated, or no documentation change is needed.

[Airbag65]

This LGTM

The final decision on this is @danieldeer 's to make however.

[danieldeer]

I would apply the following rule:

• AI simplifies code or solves bug -> acceptable
• AI adds complexity / new features -> unacceptable

Any submission, whether human or AI generated, must be validated

If AI is used for security audits, the AI output should be validated by a human before submitting. This should apply to all AI changes

If someone repeatedly submits PRs with hallucinations or unexplainable complexity, we can exclude them from contributing, and assume they are using AI

[DOT-EXECUTE]

@danieldeer I've since updated the policy. Let me know if there's any other changes needed! I'll get to them as soon as I'm able.

[AnnoDomine]

@danieldeer I've since updated the policy. Let me know if there's any other changes needed! I'll get to them as soon as I'm able.

Looks like a well-written agent markdown!
Please merge main into your branch.
Also, it would be great to add basic CLAUDE.md and GEMINI.md files pointing to AGENT.md so that AGENT.md remains our single source of truth across different LLM tools.

[danieldeer]

Different AI tools need dedicated AI .md files? @AnnoDomine

[AnnoDomine]

Different AI tools need dedicated AI .md files? @AnnoDomine

It depends on the specific tool and its configuration. For example, Claude Code looks for a CLAUDE.md by default and might ignore an AGENTS.md unless configured otherwise. While files like CLAUDE.md or .cursorrules (for Cursor) are common conventions, many tools can be configured to use a custom or shared file.
Trying to support every single tool's default file would be overkill, so we should focus on CLAUDE.md, GEMINI.md and possibly .cursorrules (i forgot this in my comment) for now. We can easily expand this later, and we've also added a note in the README pointing other LLMs to AGENTS.md.
Additionally, most AI tools apply instructions based on a clear hierarchy (global < workspace < user-defined). This means workspace configurations take precedence over global defaults.
We already have a great deal of control by managing our rules directly in the workspace-level AGENTS.md.
So it is just to ensures they aren't overlooked or overridden by default settings.

[DOT-EXECUTE]

Please merge main into your branch. Also, it would be great to add basic CLAUDE.md and GEMINI.md files pointing to AGENT.md so that AGENT.md remains our single source of truth across different LLM tools.

I have merged main into my fork. I think that tooling specific files, is outside the scope of this PR @AnnoDomine

[AnnoDomine]

Please merge main into your branch. Also, it would be great to add basic CLAUDE.md and GEMINI.md files pointing to AGENT.md so that AGENT.md remains our single source of truth across different LLM tools.

I have merged main into my fork. I think that tooling specific files, is outside the scope of this PR @AnnoDomine

Thanks for clearify.
I saw AGENTS.md and thought it would handle contributors and AI agents in the same way.
To make it more clear:
The scope is the guide contributors in which way they have to use AI and what is forbidden?!
If this is so, then AGENTS.md is the wrong file for this.

• By definition, an AGENTS.md is a document which guides an AI agent on how it has to work within this project.
  • See: https://dev.to/proflead/what-is-agentsmd-and-why-should-you-care-3bg4
• The scope is more suited for CONTRIBUTING.md then, to guide the developers on how they can use AI agents in this project.

But I would prefer having both: the developer rules in CONTRIBUTING.md, and a proper AGENTS.md instructing the AI directly. That way, we have way more control over AI usage than only guiding the developers.