)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":32968,"name":"Juan Larriba","email":"jlarriba@redhat.com","username":"jlarriba"},"change_message_id":"dbbbda5bd29754765107d9120e3f7ee939c7d75c","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"20d8de5c_7fce0a84","updated":"2026-11-03 10:03:23.000000000","message":"I was thinking about adding to the policy something that says that even if the patches can use AI to assist with writing code, getting the AI agents to talk to reviewers by posting automated comments to reviewers should not be allowed.\n\nIMHO, the patch code can be AI assisted or generated, but we should be enforcing to keep the review process strictly human-bound (even if both the reviewer and the contributor might be using AI to inquire about the code, generate a first review assessment, etc...).\n\nThat would be beneficial in at least two fronts:\n- it would discourage strictly agent-driven contributions that work on a loop where they find issues, fix them, propose the patch, read reviewer proposals, take those as AI prompts, and generate changes, all automatically.\n- it keeps the human relations that maintain the community alive, while you have interesting and valuable exchanges with other like minded people.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":10342,"name":"Jay Faulkner","display_name":"JayF","email":"jay@jvf.cc","username":"JayF","status":"youtube.com/@oss-gr / podcast.gr-oss.io"},"change_message_id":"a22dd5b39f46f62c4db67bf0367d61321ec6f08b","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"c97a6e30_5675453c","updated":"2026-11-02 20:35:14.000000000","message":"I\u0027ll note here publically as I have to Goutham in private: I am more than 100% OK with someone taking this over the finish line. I wanted a spot for us to have the discussion: mission success 😀","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"25191e02_92877cd3","updated":"2026-11-11 06:16:39.000000000","message":"Thank you for getting this started, Jay! As Mikal and others noted, not having this finalized is poor contributor experience. \n\nI apologize for the drift the next patchset will cause. I restructured and rewrote this to include several other ideas that we discussed at the Indri PTG, and elsewhere. I preserved some core ideas that Jay shared.\n\nI\u0027ll be very glad to get a re-review from everyone that has commented here.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":10342,"name":"Jay Faulkner","display_name":"JayF","email":"jay@jvf.cc","username":"JayF","status":"youtube.com/@oss-gr / podcast.gr-oss.io"},"change_message_id":"6817de34e0ee821946cef52a46553b5c1cf8bfab","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":1,"id":"edbc8a35_81a80fa5","updated":"2026-10-23 17:37:33.000000000","message":"indicating I\u0027d like this up for discussion without setting the timer on votes","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"}],"resolutions/20261023-foundation-ai-policy-interpretation.rst":[{"author":{"_account_id":10342,"name":"Jay Faulkner","display_name":"JayF","email":"jay@jvf.cc","username":"JayF","status":"youtube.com/@oss-gr / podcast.gr-oss.io"},"change_message_id":"342badcb241316acc3b893e8464f17fa05446b46","unresolved":true,"context_lines":[{"line_number":1,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":2,"context_line":" 2026-10-23 OpenInfra Foundation AI Policy Interpretation"},{"line_number":3,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":4,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"f74af8d8_7e66e8e2","line":1,"updated":"2026-10-23 17:30:28.000000000","message":"A note: I don\u0027t hold any of the opinions in this policy that strongly. I have a preference that we do not give non-human entities authorship on the change -- but the goal of posting this is to give us a place to discuss the policy and document one.\n\nI don\u0027t care as much which way this goes in terms of policy -- as long as we have something that prevents every new (or returning) contributor the opportunity to read the policy differently.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":1,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":2,"context_line":" 2026-10-23 OpenInfra Foundation AI Policy Interpretation"},{"line_number":3,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":4,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"41144dc0_ce21ef47","line":1,"in_reply_to":"f74af8d8_7e66e8e2","updated":"2026-11-11 06:16:39.000000000","message":"Thanks for getting this started. I\u0027ll attempt an update now, responding to several questions asked here, and in the community via the PTG/mailing lists and IRC discussions that I was a part of. \n\nI was also present at the Foundation\u0027s AI policy group meetings and gathered the context through which the policy was refreshed. I think we set an expectation there that the OpenStack community will discuss and adopt the policy\u0027s implementation - and that we\u0027ve some learnings to incorporate having seen this trend over teh past couple of years. Ofc we\u0027ll evolve this further as the landscape changes.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":2271,"name":"Michael Still","email":"mikal@stillhq.com","username":"mikalstill"},"change_message_id":"ff59bab444e282b486545a78ad9d63353731889f","unresolved":true,"context_lines":[{"line_number":6,"context_line":"policy for AI usage in OpenStack. However, this policy leaves some"},{"line_number":7,"context_line":"room for interpretation when implementing it. This can lead to"},{"line_number":8,"context_line":"discord in the community and be yet another barrier to new"},{"line_number":9,"context_line":"contributors."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Therefore, OpenStack Technical Committee is establishing a baseline"},{"line_number":12,"context_line":"of expected technical behaviors OpenStack contributors should"}],"source_content_type":"text/x-rst","patch_set":1,"id":"ac0bfc77_db470529","line":9,"updated":"2026-10-24 02:03:42.000000000","message":"I\u0027d tweak to something like:\n\n...another new barrier to new contributors, especially when OpenStack\u0027s preferences differ from the default configuration of many AI coding harnesses.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":6,"context_line":"policy for AI usage in OpenStack. However, this policy leaves some"},{"line_number":7,"context_line":"room for interpretation when implementing it. This can lead to"},{"line_number":8,"context_line":"discord in the community and be yet another barrier to new"},{"line_number":9,"context_line":"contributors."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Therefore, OpenStack Technical Committee is establishing a baseline"},{"line_number":12,"context_line":"of expected technical behaviors OpenStack contributors should"}],"source_content_type":"text/x-rst","patch_set":1,"id":"40feb45f_8eabcc53","line":9,"in_reply_to":"ac0bfc77_db470529","updated":"2026-11-11 06:16:39.000000000","message":"Can you share the default configuration context here, is this about harnesses forcing \"Co-Authored-By\"?\n\nI didn\u0027t want to make a reference to tools we can\u0027t control. I understand standardizing; \"Assisted-By\" seems to have gained acceptance in a lot of the open source communities.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"85993cd3e91b5bb26193bb8a79192278b931e60c","unresolved":true,"context_lines":[{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Therefore, OpenStack Technical Committee is establishing a baseline"},{"line_number":12,"context_line":"of expected technical behaviors OpenStack contributors should"},{"line_number":13,"context_line":"follow when using AI tooling on OpenStack projects."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"**Open Community \u0026 Open Design**"},{"line_number":16,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"53120894_b1f0e54b","line":13,"updated":"2026-10-28 20:26:59.000000000","message":"not sure whether to specify this here or in a dedicated section below, but I would like to have a note that individual projects (teams) may have other (like stricter) policies","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Therefore, OpenStack Technical Committee is establishing a baseline"},{"line_number":12,"context_line":"of expected technical behaviors OpenStack contributors should"},{"line_number":13,"context_line":"follow when using AI tooling on OpenStack projects."},{"line_number":14,"context_line":""},{"line_number":15,"context_line":"**Open Community \u0026 Open Design**"},{"line_number":16,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"2936d09a_5f65fc25","line":13,"in_reply_to":"53120894_b1f0e54b","updated":"2026-11-11 06:16:39.000000000","message":"ack; added","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":15,"context_line":"**Open Community \u0026 Open Design**"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"The OpenInfra policy indicates that all contributions must be"},{"line_number":18,"context_line":"compatible with the \"Four Opens\"."},{"line_number":19,"context_line":""},{"line_number":20,"context_line":"The OpenStack Technical committee interprets this to mean that"},{"line_number":21,"context_line":"contributors are required to have technical understanding of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"c8b685d3_2d256abd","line":18,"range":{"start_line":18,"start_character":20,"end_line":18,"end_character":32},"updated":"2026-11-11 06:16:39.000000000","message":"worth a link","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":17,"context_line":"The OpenInfra policy indicates that all contributions must be"},{"line_number":18,"context_line":"compatible with the \"Four Opens\"."},{"line_number":19,"context_line":""},{"line_number":20,"context_line":"The OpenStack Technical committee interprets this to mean that"},{"line_number":21,"context_line":"contributors are required to have technical understanding of"},{"line_number":22,"context_line":"code changes submitted to OpenStack projects, and should be"},{"line_number":23,"context_line":"ready to share context or design principles used if required"}],"source_content_type":"text/x-rst","patch_set":1,"id":"c2493522_02fdcaa8","line":20,"range":{"start_line":20,"start_character":34,"end_line":20,"end_character":49},"updated":"2026-11-11 06:16:39.000000000","message":"this sounds passive, and changes the meaning at least to me. I feel we need to be more assertive here; i.e., we don\u0027t interpret the Four Opens. It\u0027s our culture and we frame policies using that as the baseline.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"85993cd3e91b5bb26193bb8a79192278b931e60c","unresolved":true,"context_lines":[{"line_number":23,"context_line":"ready to share context or design principles used if required"},{"line_number":24,"context_line":"during code review. This helps us ensure that even with use"},{"line_number":25,"context_line":"of AI tooling, we maintain the ability to practice \"Open Design\""},{"line_number":26,"context_line":"in an \"Open Community\"."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"**Git Commit Trailers**"},{"line_number":29,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"b4d400a5_7178ec15","line":26,"updated":"2026-10-28 20:26:59.000000000","message":"\"ready to share\" isn\u0027t proactive enough for me, the OIF policy says:\n\n\u003e explain in comments or the commit message any prompts or background context the \n\u003e reviewers might need to fully understand the change and how much of the change\n\u003e came from the tool.\n\nso this IMO should be copied here (maybe in reworded form), to avoid reviewers having to ask for these explanations first","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":23,"context_line":"ready to share context or design principles used if required"},{"line_number":24,"context_line":"during code review. This helps us ensure that even with use"},{"line_number":25,"context_line":"of AI tooling, we maintain the ability to practice \"Open Design\""},{"line_number":26,"context_line":"in an \"Open Community\"."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"**Git Commit Trailers**"},{"line_number":29,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"99eb9b71_373e72f1","line":26,"in_reply_to":"b4d400a5_7178ec15","updated":"2026-11-11 06:16:39.000000000","message":"we have mature commit message guidance in the communkity - adding AI assistance background context, such as prompts etc there wasn\u0027t desirable iiuc from past discussions.\n\nDo you want us to require a comment instead?","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":28619,"name":"Dmitriy Rabotyagov","email":"noonedeadpunk@gmail.com","username":"noonedeadpunk"},"change_message_id":"4cc0795349ac2cc4f4c146455bf7f8713c9d3810","unresolved":true,"context_lines":[{"line_number":27,"context_line":""},{"line_number":28,"context_line":"**Git Commit Trailers**"},{"line_number":29,"context_line":""},{"line_number":30,"context_line":"The OpenInfra poilcy requires that changes be labelled in such a"},{"line_number":31,"context_line":"way as to indicate which changes used an LLM, and to indicate which"},{"line_number":32,"context_line":"human is ultimately responsible for those changes."},{"line_number":33,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"c6e81b73_2378b4d9","line":30,"range":{"start_line":30,"start_character":14,"end_line":30,"end_character":20},"updated":"2026-11-10 17:10:43.000000000","message":"policy","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"1a5060356f71109d8c4f8e37193fe69706783473","unresolved":false,"context_lines":[{"line_number":27,"context_line":""},{"line_number":28,"context_line":"**Git Commit Trailers**"},{"line_number":29,"context_line":""},{"line_number":30,"context_line":"The OpenInfra poilcy requires that changes be labelled in such a"},{"line_number":31,"context_line":"way as to indicate which changes used an LLM, and to indicate which"},{"line_number":32,"context_line":"human is ultimately responsible for those changes."},{"line_number":33,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"d6bbf03f_8c34b662","line":30,"range":{"start_line":30,"start_character":14,"end_line":30,"end_character":20},"in_reply_to":"c6e81b73_2378b4d9","updated":"2026-11-11 08:01:01.000000000","message":"Done","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":11655,"name":"Julia Kreger","email":"juliaashleykreger@gmail.com","username":"jkreger","status":"Flying to the moon with a Jetpack!"},"change_message_id":"ddeef2a322e59769b923b12904c048df4e5139bf","unresolved":true,"context_lines":[{"line_number":32,"context_line":"human is ultimately responsible for those changes."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"The OpenStack Technical Committee interprets this to mean that"},{"line_number":35,"context_line":"contributors should use git trailers in the following manner: "},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"}],"source_content_type":"text/x-rst","patch_set":1,"id":"6d9eb949_3c7b5272","line":35,"updated":"2026-10-23 17:55:49.000000000","message":"nit: whitespace *duck*","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":32,"context_line":"human is ultimately responsible for those changes."},{"line_number":33,"context_line":""},{"line_number":34,"context_line":"The OpenStack Technical Committee interprets this to mean that"},{"line_number":35,"context_line":"contributors should use git trailers in the following manner: "},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"}],"source_content_type":"text/x-rst","patch_set":1,"id":"511b4175_8d17282a","line":35,"in_reply_to":"6d9eb949_3c7b5272","updated":"2026-11-11 06:16:39.000000000","message":"Done","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":10342,"name":"Jay Faulkner","display_name":"JayF","email":"jay@jvf.cc","username":"JayF","status":"youtube.com/@oss-gr / podcast.gr-oss.io"},"change_message_id":"0b9bd7f7a50e2f15b92e5046b8677a0c8cd5e159","unresolved":true,"context_lines":[{"line_number":35,"context_line":"contributors should use git trailers in the following manner: "},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"},{"line_number":39,"context_line":"  include both the harness and model used if LLM, or the static"},{"line_number":40,"context_line":"  automation used if generated more traditionally."},{"line_number":41,"context_line":"* \"Co-Authored-By\" or \"Signed-Off-By\" trailers indicate the human"}],"source_content_type":"text/x-rst","patch_set":1,"id":"bde2b5f1_4b9ede6f","line":38,"updated":"2026-10-23 17:39:52.000000000","message":"This needs a sharper line, probably. Might make it easier to think about if we split Assisted-by/Generated-by","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":35,"context_line":"contributors should use git trailers in the following manner: "},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"},{"line_number":39,"context_line":"  include both the harness and model used if LLM, or the static"},{"line_number":40,"context_line":"  automation used if generated more traditionally."},{"line_number":41,"context_line":"* \"Co-Authored-By\" or \"Signed-Off-By\" trailers indicate the human"}],"source_content_type":"text/x-rst","patch_set":1,"id":"50222d67_6711bd6b","line":38,"in_reply_to":"202ba5d3_fb1c9381","updated":"2026-11-11 06:16:39.000000000","message":"We discussed this quite a bit over the past year - since the Foundation\u0027s policy backed away from \"Generated-By\" - i\u0027ll edit this to explicitly forbid it as an AI usage trailer - and explain why.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":2271,"name":"Michael Still","email":"mikal@stillhq.com","username":"mikalstill"},"change_message_id":"ff59bab444e282b486545a78ad9d63353731889f","unresolved":true,"context_lines":[{"line_number":35,"context_line":"contributors should use git trailers in the following manner: "},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"},{"line_number":39,"context_line":"  include both the harness and model used if LLM, or the static"},{"line_number":40,"context_line":"  automation used if generated more traditionally."},{"line_number":41,"context_line":"* \"Co-Authored-By\" or \"Signed-Off-By\" trailers indicate the human"}],"source_content_type":"text/x-rst","patch_set":1,"id":"202ba5d3_fb1c9381","line":38,"in_reply_to":"5258b9de_24c75151","updated":"2026-10-24 02:03:42.000000000","message":"I read \"generated-by\" as being \"created by a LLM with no human involvement\", whereas \"assisted-by\" reads to me as \"LLM tooling was used in the development of this patch, but this patch is not solely machine generated\". That is, I think I prefer \"assisted-by\" over \"generated-by\". I see LLM tooling as \"smart grep\" or \"smart autocomplete\". Admittedly its sometimes \"very smart autocomplete\", but its still the same as hitting tab in vscode to complete the name of a variable -- its a tool acting as a force multiplier to human effort.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":11655,"name":"Julia Kreger","email":"juliaashleykreger@gmail.com","username":"jkreger","status":"Flying to the moon with a Jetpack!"},"change_message_id":"aa1a80a55d3f804655b3394dc0b4d5b197e98b3f","unresolved":true,"context_lines":[{"line_number":35,"context_line":"contributors should use git trailers in the following manner: "},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"},{"line_number":39,"context_line":"  include both the harness and model used if LLM, or the static"},{"line_number":40,"context_line":"  automation used if generated more traditionally."},{"line_number":41,"context_line":"* \"Co-Authored-By\" or \"Signed-Off-By\" trailers indicate the human"}],"source_content_type":"text/x-rst","patch_set":1,"id":"5258b9de_24c75151","line":38,"in_reply_to":"bde2b5f1_4b9ede6f","updated":"2026-10-23 17:55:28.000000000","message":"In the policy, we backed away from generated-by, but it is also handy to understand the context of automated housekeeping commits. i.e. the pattern of opendev.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":32968,"name":"Juan Larriba","email":"jlarriba@redhat.com","username":"jlarriba"},"change_message_id":"dbbbda5bd29754765107d9120e3f7ee939c7d75c","unresolved":true,"context_lines":[{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"},{"line_number":39,"context_line":"  include both the harness and model used if LLM, or the static"},{"line_number":40,"context_line":"  automation used if generated more traditionally."},{"line_number":41,"context_line":"* \"Co-Authored-By\" or \"Signed-Off-By\" trailers indicate the human"},{"line_number":42,"context_line":"  contributor(s) who is ultimately responsible for the change and"}],"source_content_type":"text/x-rst","patch_set":1,"id":"505be34f_23a8fa5b","line":39,"range":{"start_line":39,"start_character":19,"end_line":39,"end_character":36},"updated":"2026-11-03 10:03:23.000000000","message":"One pattern I see progressively being used by myself is use a combination of models to plan, write, review and re-write a patch, which makes pretty difficult to include the \"model\" used, as it might be 6 different models contributing to the patch. This is mostly due to trying to keep the code that the LLM generates as something a human can read/maintain.\n\nIn this situation, having to fill in a model here is difficult. Could we think about not having to include such a detailed information on about what was exactly used to assist?","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":36,"context_line":""},{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"},{"line_number":39,"context_line":"  include both the harness and model used if LLM, or the static"},{"line_number":40,"context_line":"  automation used if generated more traditionally."},{"line_number":41,"context_line":"* \"Co-Authored-By\" or \"Signed-Off-By\" trailers indicate the human"},{"line_number":42,"context_line":"  contributor(s) who is ultimately responsible for the change and"}],"source_content_type":"text/x-rst","patch_set":1,"id":"eb4085e8_e9c1a6d6","line":39,"range":{"start_line":39,"start_character":19,"end_line":39,"end_character":36},"in_reply_to":"505be34f_23a8fa5b","updated":"2026-11-11 06:16:39.000000000","message":"I get that; I think the foundation policy framers wanted to know what models were involved \"materially\" - so they\u0027re relying on the contributor\u0027s judgement on which ones deserve a mention even when multiple were used for a submission.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":2271,"name":"Michael Still","email":"mikal@stillhq.com","username":"mikalstill"},"change_message_id":"ff59bab444e282b486545a78ad9d63353731889f","unresolved":true,"context_lines":[{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"},{"line_number":39,"context_line":"  include both the harness and model used if LLM, or the static"},{"line_number":40,"context_line":"  automation used if generated more traditionally."},{"line_number":41,"context_line":"* \"Co-Authored-By\" or \"Signed-Off-By\" trailers indicate the human"},{"line_number":42,"context_line":"  contributor(s) who is ultimately responsible for the change and"},{"line_number":43,"context_line":"  are indicating those human contributors agree to the DCO."}],"source_content_type":"text/x-rst","patch_set":1,"id":"be8bd08e_29a26d4e","line":40,"updated":"2026-10-24 02:03:42.000000000","message":"I like the idea of including non-LLM tooling in the requirement as you do here. I do think there\u0027s a line -- I wouldn\u0027t want to list vscode, grep, etc here, but I do think having a policy which can accomodate other significant forms of automation without modification is wise.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":37,"context_line":"* \"Assisted-by\" or \"Generated-by\" trailers must accompany any code"},{"line_number":38,"context_line":"  whose development was done using LLM. These trailers should"},{"line_number":39,"context_line":"  include both the harness and model used if LLM, or the static"},{"line_number":40,"context_line":"  automation used if generated more traditionally."},{"line_number":41,"context_line":"* \"Co-Authored-By\" or \"Signed-Off-By\" trailers indicate the human"},{"line_number":42,"context_line":"  contributor(s) who is ultimately responsible for the change and"},{"line_number":43,"context_line":"  are indicating those human contributors agree to the DCO."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3756a9b2_643453ff","line":40,"in_reply_to":"be8bd08e_29a26d4e","updated":"2026-11-11 06:16:39.000000000","message":"I think the reference here was to tooling we maintain in the community; I\u0027ll clarify this, please let me know if it\u0027s still unclear.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":11655,"name":"Julia Kreger","email":"juliaashleykreger@gmail.com","username":"jkreger","status":"Flying to the moon with a Jetpack!"},"change_message_id":"aa1a80a55d3f804655b3394dc0b4d5b197e98b3f","unresolved":true,"context_lines":[{"line_number":42,"context_line":"  contributor(s) who is ultimately responsible for the change and"},{"line_number":43,"context_line":"  are indicating those human contributors agree to the DCO."},{"line_number":44,"context_line":"  Non-human entities, such as AI or automation, are never"},{"line_number":45,"context_line":"  considered to be the author of a change."},{"line_number":46,"context_line":""},{"line_number":47,"context_line":"Multiple trailers of each type are expected in cases where multiple"},{"line_number":48,"context_line":"tools were used or multiple humans contributed to the development of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"8be72dd3_3517ee1a","line":45,"updated":"2026-10-23 17:55:28.000000000","message":"Note: and this was mentioned by opendev personnel on irc: If openstack chooses to enforce it, it would need to be in CI job configuration to deliverables to enforce whatever combo and precise interpretation of enforcement is desired.\n\nThe key major difference is we moved from CLA modeling to DCO modeling, so each committer may be good and assert the state of the code as when they touched it, but they also can\u0027t speak for it as it evolves as the commit is amended, i.e. they are \"co-authors\" but not submitting directly so maybe not \"Signed-Off-By\". Its an interpretation thing much like this entire discussion as well.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":42,"context_line":"  contributor(s) who is ultimately responsible for the change and"},{"line_number":43,"context_line":"  are indicating those human contributors agree to the DCO."},{"line_number":44,"context_line":"  Non-human entities, such as AI or automation, are never"},{"line_number":45,"context_line":"  considered to be the author of a change."},{"line_number":46,"context_line":""},{"line_number":47,"context_line":"Multiple trailers of each type are expected in cases where multiple"},{"line_number":48,"context_line":"tools were used or multiple humans contributed to the development of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"54897de0_194062c7","line":45,"in_reply_to":"2c7edf40_b19c299c","updated":"2026-11-11 06:16:39.000000000","message":"+1; i\u0027ll repeat this argument as to why an LLM can\u0027t be a co-author. \n\nI don\u0027t know the answer to enforcement though beyond human code review.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":2271,"name":"Michael Still","email":"mikal@stillhq.com","username":"mikalstill"},"change_message_id":"ff59bab444e282b486545a78ad9d63353731889f","unresolved":true,"context_lines":[{"line_number":42,"context_line":"  contributor(s) who is ultimately responsible for the change and"},{"line_number":43,"context_line":"  are indicating those human contributors agree to the DCO."},{"line_number":44,"context_line":"  Non-human entities, such as AI or automation, are never"},{"line_number":45,"context_line":"  considered to be the author of a change."},{"line_number":46,"context_line":""},{"line_number":47,"context_line":"Multiple trailers of each type are expected in cases where multiple"},{"line_number":48,"context_line":"tools were used or multiple humans contributed to the development of"}],"source_content_type":"text/x-rst","patch_set":1,"id":"2c7edf40_b19c299c","line":45,"in_reply_to":"8be72dd3_3517ee1a","updated":"2026-10-24 02:03:42.000000000","message":"I haven\u0027t really been paying attention to the CLA / DCO thing apart from attempting to comply with it as best as I can, but it seems to me the main issue is that if the DCO is seen as a complete replacement for the CLA then it is effectively a copyright assignment. That is, a \"co-authored-by\" is the submitter asserting that the following authors consent to their work being licensed under apache2 irrevocably. Implicit in that is the assertion that they are _authorized_ to gift that code in that way I suppose.\n\nSo by that argument, a LLM (or any other automation) does not belong in a \"co-authored-by\" because tools cannot currently enter into contracts and the copyright assignment mechanism is inherently contractual.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":2271,"name":"Michael Still","email":"mikal@stillhq.com","username":"mikalstill"},"change_message_id":"ff59bab444e282b486545a78ad9d63353731889f","unresolved":true,"context_lines":[{"line_number":46,"context_line":""},{"line_number":47,"context_line":"Multiple trailers of each type are expected in cases where multiple"},{"line_number":48,"context_line":"tools were used or multiple humans contributed to the development of"},{"line_number":49,"context_line":"the patch."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":".. _OpenInfra Foundation AI policy: https://openinfra.org/legal/ai-policy"}],"source_content_type":"text/x-rst","patch_set":1,"id":"59de6206_8be04b3f","line":49,"updated":"2026-10-24 02:03:42.000000000","message":"I think this also needs to clarify _how_much_ contribution triggers the \"assisted-by\" requirement. Does `grep` usage require it? If not, then if I use a model for \"grep-like\" things, then why would it require attribution? It seems to me that it matters only if new code is created by the tool, not merely lookups of existing code?\n\nBy that argument, I think we should be requiring \"assisted-by\" stanzas for say protobuf compilers. That would presumably include the version of the compiler as the equivalent to model / effort. I think that\u0027s a good idea, because it would help us track if we were exposed to some sort of systemic bug in a protobuf compiler if one was discovered, in the same way that we might be exposed to some systemic error in a specific model\u0027s output.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":32968,"name":"Juan Larriba","email":"jlarriba@redhat.com","username":"jlarriba"},"change_message_id":"dbbbda5bd29754765107d9120e3f7ee939c7d75c","unresolved":true,"context_lines":[{"line_number":46,"context_line":""},{"line_number":47,"context_line":"Multiple trailers of each type are expected in cases where multiple"},{"line_number":48,"context_line":"tools were used or multiple humans contributed to the development of"},{"line_number":49,"context_line":"the patch."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":".. _OpenInfra Foundation AI policy: https://openinfra.org/legal/ai-policy"}],"source_content_type":"text/x-rst","patch_set":1,"id":"ee702ae0_514a1ea7","line":49,"in_reply_to":"59de6206_8be04b3f","updated":"2026-11-03 10:03:23.000000000","message":"I agree with Michael, I dont see this as really useful.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"3d8ea51ab655eda39524eb5cea520269de2153f8","unresolved":false,"context_lines":[{"line_number":46,"context_line":""},{"line_number":47,"context_line":"Multiple trailers of each type are expected in cases where multiple"},{"line_number":48,"context_line":"tools were used or multiple humans contributed to the development of"},{"line_number":49,"context_line":"the patch."},{"line_number":50,"context_line":""},{"line_number":51,"context_line":".. _OpenInfra Foundation AI policy: https://openinfra.org/legal/ai-policy"}],"source_content_type":"text/x-rst","patch_set":1,"id":"bc3d1b3a_e83f6245","line":49,"in_reply_to":"ee702ae0_514a1ea7","updated":"2026-11-11 06:16:39.000000000","message":"I reworded this - and left the extent of disclosure to the contributors\u0027 judgement. There\u0027s no one-size fits all unless we tell them give us _everything_ you\u0027ve used for the contribution.","commit_id":"54fa6329559b61a58109936b282cb6773b1dc45c"},{"author":{"_account_id":6914,"name":"Tatiana Ovchinnikova","email":"t.v.ovtchinnikova@gmail.com","username":"tmazur"},"change_message_id":"73ee829ef2d1d050df58b422c5e3e99f068856a6","unresolved":true,"context_lines":[{"line_number":88,"context_line":""},{"line_number":89,"context_line":"* A human is accountable for every contribution that they"},{"line_number":90,"context_line":"  sign-off under the Developer Certificate of Origin (DCO)."},{"line_number":91,"context_line":"  They must stand behind the contents of the contribution."},{"line_number":92,"context_line":"* A human understands their contribution and can explain all"},{"line_number":93,"context_line":"  aspects of it during code review, including the design"},{"line_number":94,"context_line":"  principles, the context, and any prompts or background a"}],"source_content_type":"text/x-rst","patch_set":6,"id":"79f5afcf_f1a489f5","line":91,"updated":"2026-11-12 15:38:25.000000000","message":"Could we explicitly state that contributions must be tested before being marked as ready for review? While this isn\u0027t strictly an AI issue, we\u0027ve seen a recent uptick in AI-assisted contributions that fail basic runs, either not fixing the original issue or introducing new ones. That\u0027s completely fine for a WIP, but marking broken code as ready for review costs maintainers valuable time spinning up test environments just to find out it doesn\u0027t work.","commit_id":"f57470ca7b2969fc4f348c306fdcff015599b0f9"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"f87e5de2c3fa93b34998311edd3fb486726c8af1","unresolved":true,"context_lines":[{"line_number":88,"context_line":""},{"line_number":89,"context_line":"* A human is accountable for every contribution that they"},{"line_number":90,"context_line":"  sign-off under the Developer Certificate of Origin (DCO)."},{"line_number":91,"context_line":"  They must stand behind the contents of the contribution."},{"line_number":92,"context_line":"* A human understands their contribution and can explain all"},{"line_number":93,"context_line":"  aspects of it during code review, including the design"},{"line_number":94,"context_line":"  principles, the context, and any prompts or background a"}],"source_content_type":"text/x-rst","patch_set":6,"id":"2e8776fe_7a78255e","line":91,"in_reply_to":"1b1e04f7_8a5b0f55","updated":"2026-11-13 06:18:10.000000000","message":"If a contributor does this once, or a few times, we would tell them via the review that they\u0027re wasting CI cycles and it\u0027s reflecting poorly on their ability to gain trust in the community. Typically new contributors need to be told this a few times before they get it. CI is a shared resource and cannot be abused - I don\u0027t know if this feels worse with AI tools being used.\n\nIf AI assisted contributions are unable to meet this bar, it\u0027s fair to refuse such changes. I think we can report the gerrit abuse to infra so they can ban the account. I\u0027ll explore how we can codify wilful or mindless CI abuse, it seems in a similar vein to what Dan added below.","commit_id":"f57470ca7b2969fc4f348c306fdcff015599b0f9"},{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"dd599997938ab261f44b3fb23a0a9ef3951d3c76","unresolved":true,"context_lines":[{"line_number":88,"context_line":""},{"line_number":89,"context_line":"* A human is accountable for every contribution that they"},{"line_number":90,"context_line":"  sign-off under the Developer Certificate of Origin (DCO)."},{"line_number":91,"context_line":"  They must stand behind the contents of the contribution."},{"line_number":92,"context_line":"* A human understands their contribution and can explain all"},{"line_number":93,"context_line":"  aspects of it during code review, including the design"},{"line_number":94,"context_line":"  principles, the context, and any prompts or background a"}],"source_content_type":"text/x-rst","patch_set":6,"id":"1b1e04f7_8a5b0f55","line":91,"in_reply_to":"79f5afcf_f1a489f5","updated":"2026-11-12 20:32:18.000000000","message":"I\u0027m not sure about this, I wouldn\u0027t require it for human contributions. not everyone has the resources and skill necessary for that, and also we do have the zuul CI doing the testing for us?","commit_id":"f57470ca7b2969fc4f348c306fdcff015599b0f9"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"6c33e4744610ca302801c1967fef0747009eabd9","unresolved":true,"context_lines":[{"line_number":102,"context_line":"* Project teams reject contributions that are completely"},{"line_number":103,"context_line":"  autonomously generated via AI tooling, except in cases where"},{"line_number":104,"context_line":"  the contribution comes from well documented, community-run"},{"line_number":105,"context_line":"  automated processes."},{"line_number":106,"context_line":""},{"line_number":107,"context_line":"Disclosure via contribution commit messages"},{"line_number":108,"context_line":"-------------------------------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"88a21dde_66f15365","line":105,"updated":"2026-11-11 16:51:57.000000000","message":"This would also preclude cases where a prompt generated code *and* autonomously submitted it to gerrit without human review first right? And presumably that also precludes even crazier stuff like \"I left my agent running overnight and let it automatically \u0027fix\u0027 things in my submission after it sees zuul output\" right?\n\nIf so, could we specifically say that here? Only because I was shocked to find out some people are doing things like this. Even with a WIP tag, I definitely do not like this behavior for submissions against major projects, personally.","commit_id":"f57470ca7b2969fc4f348c306fdcff015599b0f9"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"f3530d098cd03ef17151d7288fbd6eca252a164c","unresolved":true,"context_lines":[{"line_number":102,"context_line":"* Project teams reject contributions that are completely"},{"line_number":103,"context_line":"  autonomously generated via AI tooling, except in cases where"},{"line_number":104,"context_line":"  the contribution comes from well documented, community-run"},{"line_number":105,"context_line":"  automated processes."},{"line_number":106,"context_line":""},{"line_number":107,"context_line":"Disclosure via contribution commit messages"},{"line_number":108,"context_line":"-------------------------------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"eb25bf8e_70692fa6","line":105,"in_reply_to":"6be49e3e_069fbdf2","updated":"2026-11-13 18:29:25.000000000","message":"Yeah, I hate to even say this, but it would be trivial to tell your agent to just submit over and over and parse the test results overnight while you sleep instead of using your own cores/electricity to run unit tests locally.","commit_id":"f57470ca7b2969fc4f348c306fdcff015599b0f9"},{"author":{"_account_id":16643,"name":"Goutham Pacha Ravi","email":"gouthampravi@gmail.com","username":"gouthamr"},"change_message_id":"f87e5de2c3fa93b34998311edd3fb486726c8af1","unresolved":true,"context_lines":[{"line_number":102,"context_line":"* Project teams reject contributions that are completely"},{"line_number":103,"context_line":"  autonomously generated via AI tooling, except in cases where"},{"line_number":104,"context_line":"  the contribution comes from well documented, community-run"},{"line_number":105,"context_line":"  automated processes."},{"line_number":106,"context_line":""},{"line_number":107,"context_line":"Disclosure via contribution commit messages"},{"line_number":108,"context_line":"-------------------------------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"6be49e3e_069fbdf2","line":105,"in_reply_to":"88a21dde_66f15365","updated":"2026-11-13 06:18:10.000000000","message":"Ack, CI is a shared/community resource and using it to iterate over changes autonomously is unfair/abusive. I\u0027ll figure out a way to encode this.","commit_id":"f57470ca7b2969fc4f348c306fdcff015599b0f9"}]}
