菜单
Run a pre-PR or pre-push quality gate by inspecting the branch diff, tightening touched public API docstrings/types, running validation, and summarizing residual risk.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
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.
| name | pr-gate |
| description | Run a pre-PR or pre-push quality gate by inspecting the branch diff, tightening touched public API docstrings/types, running validation, and summarizing residual risk. |
Prepare a branch for PR or push:
Identify these before editing:
Default validation is pre-commit run --all-files and pytest -q only when the
repo and environment make those reasonable. Otherwise use the strongest targeted
checks and explain the scope.
Check branch and workspace state, then inspect the chosen range:
git branch --show-current
git status --short
git diff --name-status <base>...HEAD
Do not overwrite unrelated dirty files.
Read critical diffs, not only filenames. Focus on public API changes, runtime behavior, config contracts, tests, docs, and generated artifacts.
Read references/docstring-and-typing.md when touched APIs need cleanup.
Only edit files already in scope. Keep docstrings and type hints aligned with
actual behavior; avoid filler comments and broad Any unless justified.
Run formatting/lint first, then tests. Fix failures caused by the current change set. If a gate is blocked, record the blocker and residual risk.
List findings by severity, validation commands and outcomes, and any remaining unvalidated paths.
severity | file:line | issue | recommendation.command | result.references/docstring-and-typing.md: cleanup scope and examples.