[Klaud Cold][DO NOT MERGE] Test signoff-verify Check 5 with stale checklist template

[functionstackx]

Summary

• Trivial README touch (mirrored in README_zh per the bilingual rule) to host a CODEOWNER sign-off that uses a STALE pre-[merging June 30 at 4pm PT] making this an hard guideline & enforcing consistent reviews on upstream sglang/vllm docker repo to PR CheckList #1956 checklist template — missing the upstream-Docker-image item and the no-benchmark-hacks item.
• NEGATIVE test for the updated codeowner-signoff-verify workflow ([Klaud Cold] Align codeowner-signoff-verify with latest PR_REVIEW_CHECKLIST.md #2015): Check 5 should FAIL naming the missing items.
• Will be closed without merging once verified.

🤖 Generated with Claude Code

[functionstackx]

As a PR reviewer and CODEOWNER, I have reviewed this and have:

• Verified that as of the moment of typing this, this is the latest version of PR_REVIEW_CHECKLIST.md
• Verified that the general code quality meets the InferenceX standard and does not make the code quality any worse.
• Verified that this PR has passed PR validation. Please link to GitHub Action workflow that shows this.
• Verified that this PR passes evals. Please link to GitHub Action workflow that shows this.
• Verified that speculative decoding PRs uses chat templates to align the AL distribution to real world
• If an company claims that they support vLLM/SGLang as first class LLM inference engines on their hardware, I have have verified that the respective vLLM/SGLang submission has been made before additional frameworks (TRT-LLM, ATOM, etc.). The only exceptions are for new hardware, such as MI455X UALoE72, Vera Rubin NVL72, Rubin NVL8, etc., and for new model architectures where there is an actual reason why vLLM/SGLang does not fundamentally support them yet.
• Verified that the single-node recipes are similar to the official vLLM recipes and/or theSGLang cookbook:
  • If they are not, I have verified that a PR has been opened in vLLM recipe repo or SGLang repo and linked it below in the additional detail section:
• If any of the above criteria cannot reasonably be satisfied, I have provided additional reasoning below.

Additional detail section:

• Docs-only test PR.

Signed: functionstackx

[Klaud-Cold]

@functionstackx FAIL — merge blocked: no passing sweep/eval on any commit in this PR, no authorized /reuse-sweep-run posted, and the sign-off uses a stale checklist template missing two current items.

• Check 0 (CODEOWNER): PASS — changed paths (README.md, README_zh.md) match only the catch-all * @InferenceX/core; signer is an org member and no specific-owner line is violated.
• Check 1 (sweep on in-PR commit): FAIL — no passing sweep/eval was found on any commit in this PR. The sole commit 3849f84 has no single-node */ or eval / check-runs at all (only CodeQL/analysis jobs).
• Check 2 (evals pass): FAIL — blocked by Check 1; no eval run exists to verify accuracy against.
• Check 3 (recipe linked and complete): N/A — docs-only diff (README comment lines); no benchmarks/** or master-config changes, so no recipe applies.
• Check 4 (reuse command): FAIL — no authorized /reuse-sweep-run command has been posted on this PR; an authorized maintainer must comment it before merge via reuse.
• Check 5 (latest checklist template): FAIL — sign-off is missing two items from the current docs/PR_REVIEW_CHECKLIST.md: (1) model architecture not changed via benchmark hacks (no FLOPs-reducing overrides like --hf-overrides), and (2) vLLM/SGLang submissions use upstream vllm/lmsysorg Docker repos.
• Check 6 (upstream images / engine-first): N/A — no master-config entries changed.
• Check 7 (no architecture hacks): N/A — no server args or benchmark scripts touched.
• Check 8 (spec-decode chat templates): N/A — no speculative-decoding changes.

Note: PR is titled DO NOT MERGE and self-describes as a negative test of this workflow; the failures above are consistent with that intent.

[functionstackx]

Negative test confirmed — Check 5 correctly FAILed on the stale template, naming both missing items. Closing.