メニュー
Triage reviewer feedback on an open PR, implement scoped fixes, rerun validation, and prepare reviewer-ready responses tied to comments, changed files, and validation evidence.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
Use and extend the MVP-Engine MLLM data kit with explicit specs, handlers, and sample objects. Use for recipe data wiring, MLLMDataSpec construction, source resample/resolve_refs policy, schema/media/tokenization handlers, packing, guards, collation, dataloader setup, Qwen VL data support, and multimodal data extensions.
Add, review, update, and validate recipe-local sequence-parallel plans that reuse the tensor-parallel mesh in mvp-engine models.
Add, review, update, and validate recipe-local tensor-parallel plans, TP metadata postprocessors, and compatible mesh config for mvp-engine models.
Add, review, update, and validate VLM/MLLM data pipelines using the current spec/handler-based MLLMDataKit design. Use for raw schema normalization, MLLMDataSpec wiring, source resample/resolve_refs policy, media extensions, processor setup, packing, guards, collation, step estimation, and recipe integration.
Add, review, update, and validate VLM/MLLM packing behavior around the current MLLMDataKit design, including MLLMPackingSpec, MLLMPackingAssembler, MLLMPack metadata, block-causal masks, model-specific packed input preparation, token accounting, and step estimation.
Use LigerKernelKit for reusable Liger Kernel integration before model construction, covering official model-family dispatch, custom-model symbol patching via LigerPatch, module selection validation, and loss-kernel guards.
SOC 職業分類に基づく
| name | pr-feedback |
| description | Triage reviewer feedback on an open PR, implement scoped fixes, rerun validation, and prepare reviewer-ready responses tied to comments, changed files, and validation evidence. |
Resolve reviewer feedback without broadening the PR:
Identify these before editing:
Ask the user only when reviewer intent or permission to mutate/push is unclear.
Gather comments and map each to:
Group comments that require the same edit. Keep explanation-only comments separate from code changes.
If comments conflict, identify the conflict before editing and ask for a decision only when the repo context cannot resolve it.
Make the smallest correct change for each actionable comment. Update docstrings, typing, tests, or docs only when the behavior contract changed or the reviewer asked for it.
Avoid unrelated cleanup and do not revert local changes you did not make.
Run targeted tests for touched behavior first, then broader gates when feasible. If full validation is too expensive or blocked, state the exact remaining gap.
For every comment, prepare:
Read references/feedback-checklist.md when the PR changes skills.
comment id/topic | fixed/clarified/deferred/pending.command | result.references/feedback-checklist.md: skill-PR feedback dimensions.../pr-gate/references/docstring-and-typing.md: docstring/type cleanup rules.