PR #8565
llm: pr-conversation -> reviewer (The author asked for guidance on the preferred testing/implementation approach, so the next action is for a reviewer/maintainer to जवाब/decide.)

PR #8564
llm: pr-conversation -> reviewer (The reviewer approved but explicitly deferred to another reviewer to weigh in, so the next action is on the reviewer/maintainer side rather than the author.)

PR #8548
llm: pr-conversation -> reviewer (The latest comment is from a reviewer asking the approvers to review the proposed direction and share thoughts, so the next move is on the reviewer/maintainer side.)

PR #8545
llm: PRRT_kwDOCkv3g86N_pdf -> none (The author responded to the reviewer’s naming suggestion and explicitly chose not to change it, so no further follow-up is needed.)
llm: pr-conversation -> none (The latest comment is an approval/thanks with no requested follow-up, and the earlier author replies already addressed the review points.)

PR #8527
llm: pr-conversation -> reviewer (The latest comment is from the author and it answers the prior concern with updated analysis; the ball is back with the reviewer to acknowledge or continue review.)

PR #8497
llm: pr-conversation -> reviewer (The author says the PR is ready for review after making the change, so the next action is on a reviewer/maintainer to review it.)

PR #8496
llm: pr-conversation -> reviewer (The PR author’s latest comment says the change is ready for review and CI is green, so the next action is for a reviewer/maintainer to review or merge.)

PR #8495
llm: pr-conversation -> reviewer (The author says the PR is ready for review and is inviting reviewer input on whether to add coverage, so the next move is on the reviewer.)

PR #8494
llm: pr-conversation -> reviewer (The author says the PR is ready for review and CI is green, so the next action is for a reviewer/maintainer to review it.)

PR #8493
llm: pr-conversation -> reviewer (The author’s latest comment says the PR is ready for review and CI is green, so the ball is with a reviewer/maintainer.)

PR #8467
llm: pr-conversation -> reviewer (The reviewer asked for motivation, and the author answered with the rationale and linked issue; the ball is back with the reviewer to acknowledge or continue review.)

PR #8464
llm: PRRT_kwDOCkv3g86MPsyw -> author (The latest reviewer comment requests a change to enforce non-empty IDs at construction time, so the PR author needs to implement or जवाब back.)

PR #8450
llm: pr-conversation -> none (The author only reports that they restored the branch to a previous state; there’s no explicit request or pending follow-up in the thread.)

PR #8446
llm: PRRT_kwDOCkv3g86KBUh4 -> author (A reviewer asked a direct question (“Why two methods?”), so the author needs to जवाब/respond or adjust the code.)
llm: PRRT_kwDOCkv3g86KBWSq -> author (A reviewer left a code change suggestion and the thread is still unresolved, so the PR author needs to respond by applying or addressing it.)

PR #8428
llm: PRRT_kwDOCkv3g86FNGwS -> author (A reviewer asked for the rationale behind the chosen number, so the author needs to जवाब/respond with an explanation or update.)
llm: PRRT_kwDOCkv3g86JKWNr -> author (A reviewer asked a code question/suggestion and there is no follow-up yet, so the author needs to respond or update the PR.)

PR #8407
llm: PRRT_kwDOCkv3g86CMfQS -> none (The reviewer asked for explanatory comments, and the author replied that they added them, so the request is addressed and no further follow-up is implied.)
llm: PRRT_kwDOCkv3g86CMnfF -> reviewer (The reviewer asked why it changed; the author explained the rationale, so the ball is back with the reviewer to accept or continue the review.)
llm: pr-conversation -> author (A reviewer’s last comment asks the author to try `@CompileStatic` and questions whether reflection is necessary, so the ball is with the author to respond or update the PR.)

PR #8364
llm: PRRT_kwDOCkv3g86BhQsA -> author (The latest reviewer comment requests a code change to reduce allocation overhead, so the PR author needs to update the thread.)
llm: PRRT_kwDOCkv3g86BhVsZ -> author (A reviewer指出 collision handling should also account for normalized attributes from other sources, so the PR author needs to respond and update the implementation.)
llm: pr-conversation -> author (The author says the PR is parked for #8346 and that they will rebase, apply the requested changes, and re-request review once that settles, so the next action is on the author.)

PR #8349
llm: pr-conversation -> reviewer (The latest comment is from the author responding to the review points and asking for clarification on the `sampledThread` concern, so the reviewer has the next turn.)

PR #8240
llm: pr-conversation -> author (The author’s latest comment says they will keep investigating why the benchmark metrics are zero, so the ball is still with the author.)

PR #8197
llm: pr-conversation -> external (The reviewer asked to wait for spec discussion, and the author’s latest comment just points to the external specification issue, so the next step is outside this repository.)

PR #8164
llm: PRRT_kwDOCkv3g85z-n0C -> none (The latest reviewer comment is a brief approval of the proposed convention, which closes the thread with no further action needed.)

PR #8076
llm: pr-conversation -> author (The latest comment is from the author and says they will still test and validate this week, so the ball remains with the author.)

PR #7763
llm: pr-conversation -> reviewer (The reviewer asked why, and the author answered with an explanation; the ball is back with the reviewer to acknowledge or continue the review.)

PR #7741
llm: pr-conversation -> author (The latest comment is a reviewer suggestion with an explicit request for the author to evaluate it, so the author needs to respond or act.)