Warn on missing NVIDIA recipe overview entries / 对 NVIDIA 配方概览缺失条目发出警告

[functionstackx]

Add advisory Check 16 to the AI CODEOWNER reviewer. For NVIDIA model/framework pairs documented in merged/published vLLM Recipes or the SGLang Cookbook, it checks NVIDIA's official overview and emits a visible ❗❗ WARNING ❗❗ when the pair is missing.

The check covers direct and Dynamo-backed vLLM/SGLang deployments, preserves model versions/variants, and excludes all non-NVIDIA hardware, including AMD MI300/MI325/MI355/MI455, TPUv7/TPUv6e, Trainium3, and SambaNova SN50. Inaccessible or ambiguous evidence is reported as unverified, not as a confirmed missing listing.

Update the verdict publisher and checklist documentation for the additional advisory row. Existing recipe-completeness failures remain failures; this warning does not trigger Pareto-reviewer mentions. No benchmark configs, recipes, or performance behavior change.

AI model disclosure

• Model/version: GPT-6 Astra Ultrafast (runtime-provided model name).
• Role: implementation, documentation, local verification, and PR preparation. No delegated models.

PR checklist and change type

Type of Change

• Bug fix
• New feature
• Configuration change
• Documentation update
• Other

Checklist

• I have completed the AI model disclosure and kept it current.
• I have tested my changes locally.
• I have updated documentation if necessary.
• Performance changelog requirement assessed: N/A, no performance-affecting or recipe changes.
• Primary sweep label requirement assessed: N/A, no performance-affecting or recipe changes.
• Reuse command requirement assessed: N/A, no benchmark sweep artifacts to reuse.

中文

为 AI CODEOWNER 审阅器新增建议性检查 Check 16。对于已在合并或发布的 vLLM Recipes 或 SGLang Cookbook 中提供文档的 NVIDIA 模型/框架组合，检查 NVIDIA 官方概览 是否列出该组合；若缺失，则显示 ❗❗ WARNING ❗❗。

检查涵盖直接部署和基于 Dynamo 的 vLLM/SGLang 部署，区分模型版本及变体，并排除所有非 NVIDIA 硬件，包括 AMD MI300/MI325/MI355/MI455、TPUv7/TPUv6e、Trainium3 和 SambaNova SN50。无法访问或存在歧义的证据会标记为未验证，不会据此断言条目缺失。

同步更新结果发布器和审阅清单文档。现有配方完整性检查的失败结果保持不变；新警告不会触发 Pareto 审阅者提醒。本 PR 不修改基准配置、配方或性能行为，无需性能变更记录、sweep 标签或结果复用命令。

AI 模型：GPT-6 Astra Ultrafast（运行环境提供的名称），用于实现、文档、本地验证及 PR 准备，未委派其他模型。改动类型为新功能及文档更新；已完成本地检查和文档更新。

[github-actions]

inferencex-e2e/infx diff: +70 -14

Generally most InferenceX submissions don't need to add changes to the infx package. On the off chance they do, generally the diff is small. Please prompt your agent to clean your PR to keep your diff to only necessary & minimal changes too. We have an increased code quality bar for changes to the infx package.

中文

大多数 InferenceX 提交通常不需要修改 infx 包。即使确实需要，改动通常也很小。请让你的 agent 清理 PR，使改动只包含必要且最小的内容。infx 包的改动有更高的代码质量要求。

[claude]

LGTM — small, well-scoped addition of an advisory Check 16 that follows the existing Check 14 pattern throughout. Reviewed the prompt doc renumbering (0-15 → 0-16), the signoff_publish.py changes to check_statuses/verdict_body (multi-notice collection, still only WARN-eligible for checks 4/14/15/16, still requires the full 0-16 range), and the corresponding test updates; also checked the flagged long test line (332, 102 chars) against inferencex-e2e/infx/ruff.toml — tests is excluded from Ruff and E501 is globally ignored, so it will not break CI lint.

Extended reasoning...

The diff adds a new advisory Check 16 (NVIDIA overview coverage) across the verifier prompt, checklist doc, publisher workflow, and its tests, closely mirroring the existing Check 14/15 handling; it touches no auth, crypto, or data-exposure surface, only string construction and status-dict bookkeeping. check_statuses and verdict_body changes are small, symmetric extensions of existing logic, and the new tests in test_signoff_publish.py exercise real end-to-end behavior through the publish fixture rather than reimplementing or pinning source text. The one previously-flagged candidate (a test line exceeding 100 chars) was independently confirmed to be a non-issue because ruff.toml excludes the tests directory and ignores E501 entirely, so it will not fail CI.