docs: Update the AI code guidelines

CONTRIBUTING.md

@@ -8,13 +8,9 @@ To contribute, fork this repository to create your own copy that you can clone l
 Prefer opening fewer high quality Pull Requests over many low quality ones. Prioritize polishing existing Pull Requests instead of creating new ones. Do not open Pull Requests as draft unless there are specific requirements that justify it, because draft Pull Requests are generally not looked at and pollute the Pull Request section.
 
 
-## AI code generation guidelines
+## LLM contribution policy
 
-Creating changes with LLM generated code is generally allowed. The author is responsible for verifying that all generated code is human readable, maintainable and logically correct. Furthermore, all generated code needs to be tested and verified. The author is not allowed to outsource the polishing of the generated code to the human code reviewers. In a Pull Request, generated code needs to be announced as such and to what extent it was polished by human intervention.
-
-### New code contributors and AI generated code
-
-New contributors are discouraged from submitting Pull Requests with thousands of lines changed or added with the help of LLMs, because human code reviewers cannot attend such volumes at the risk of wasting precious time with potentially poorly generated code.
+Ensure that you have read and understood the [LLM Contribution Policy](LLM_POLICY.md).
 
 
 ## Code guidelines

LLM_POLICY.md

@@ -0,0 +1,55 @@
+# LLM Contribution Policy
+
+Today, large language model (LLM)-based tools are being deployed in many coding environments. Some of the "lower-end" features like keyword-driven code completion can be useful aids; other tools make it very easy to generate contributions without the submitter fully understanding how the code works, or the consequences of specific implementation choices. These latter forms of LLM assistance cause major overhead for project maintainers.
+
+To protect the quality of the project and the sanity of our contributors, we have adopted the following policy regarding LLM-generated content.
+
+## 1. Accountability
+
+**The human contributor is the sole party responsible for the contribution.**
+
+If you submit a pull request that includes LLM-generated code, documentation, or comments:
+
+- You must fully understand every line of code in the submission.
+- You must be able to explain the "why" behind the implementation during the review process.
+- You are responsible for verifying that the code is human-readable, maintainable and logically correct.
+- All code needs to be tested and verified.
+
+"The LLM generated it and it works for me" is never an acceptable answer to a reviewer's question. You must write review responses yourself. LLM tools may assist with translation, meaning-preserving edits, or formatting, but you must verify the final wording. If unsure of an answer, say so rather than presenting an LLM's output as your own.
+
+Autonomous LLM agents may not submit pull requests, nor comments in other pull requests, issues or discussions.
+
+## 2. Disclosure
+
+If an LLM was used to generate a significant portion of your contribution (i.e. beyond simple autocomplete), we require you to **disclose it** in the pull request description. Note that you should not use an LLM to create the pull request description itself (unless you’ve used it for translation) - as discussed above, we expect pull requests to be submitted by humans.
+
+In a pull request, generated code needs to be announced as such and to what extent it was polished by human intervention.
+
+## 3. Intentionality
+
+We do not accept pull requests and issues that result from running an LLM tool over the codebase to find improvements without prior context or alignment with the project.
+
+- **Focused changes:** Do not submit pull requests that perform broad refactoring or cleanup suggested by an LLM unless a maintainer specifically requests it.
+- **Design First:** For any non-trivial change, we strongly recommend opening an **Issue** or **Discussion** first, or reach out to the **Discord**. Pull requests that arrive out of the blue with significant LLM-generated logic that doesn't align with our roadmap or architecture will be closed.
+- **Quality over Quantity:** We value one thoughtful, manually crafted pull request over ten LLM-assisted fixes for nonexistent or trivial issues.
+
+New contributors are discouraged from submitting pull requests with thousands of lines changed or added with the help of LLMs, because human code reviewers cannot attend such volumes at the risk of wasting precious time with potentially poorly generated code.
+
+## 4. Prohibited Uses
+
+The following are strictly prohibited and will result in immediate closure of a pull request or issue and potentially a block from the organization:
+
+- **Automated Pull Request Descriptions:** Using an LLM to write pull request descriptions that are vague, overly elaborate, or fail to accurately describe the technical changes. We want to hear from *you* - the developer - why this change matters (see points 1 & 2).
+- **Unvetted Boilerplate:** Submitting large blocks of LLM-generated boilerplate that haven't been trimmed to what's actually necessary. If you don't understand what the code does, don't submit the pull request.
+- **Hallucinated Features:** Submitting pull requests for features or bug fixes that don't exist, based on LLM hallucinations about the project's capabilities.
+
+## 5. Enforcement
+
+TheSuperHackers project maintainers reserve the right to close any pull request that appears to be a low-effort LLM-assisted contribution, without providing a detailed technical critique. If a maintainer suspects you do not understand your pull request, it will be closed immediately.
+
+Cases of repeated violations of these (or any of our other contributor guidelines) could result in a ban from our repositories.
+
+### Acknowledgement
+
+This policy was written by humans, based on the work in the [Mastodon AI Policy](https://github.com/mastodon/.github/blob/main/AI_POLICY.md), the [CloudNativePG AI Policy](https://github.com/cloudnative-pg/governance/blob/main/AI_POLICY.md), the [Ghostty AI Policy](https://github.com/ghostty-org/ghostty/blob/main/AI_POLICY.md), and the Linux Foundation's [Generative AI Policy](https://www.linuxfoundation.org/legal/generative-ai).
+