BBuf/vllm-qwen35-optimization
skills/model-optimization/vllm/vllm-qwen35-optimization/SKILL.md
PR-backed optimization manual for Qwen3.5 in vLLM. Use when an engineer needs to audit, debug, extend, or document Qwen3.5 dense / MoE / GDN runtime, MTP, FP8 and NVFP4 quantization, LoRA, and Eagle3 in vLLM.
$ install --global
skillsauthnpx skillsauth add BBuf/AI-Infra-Auto-Driven-SKILLS vllm-qwen35-optimizationInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: May 2, 2026, 2:45 AM331.7s2 files scanned
SKILL.md
- name:
- vllm-qwen35-optimization
- description:
- PR-backed optimization manual for Qwen3.5 in vLLM. Use when an engineer needs to audit, debug, extend, or document Qwen3.5 dense / MoE / GDN runtime, MTP, FP8 and NVFP4 quantization, LoRA, and Eagle3 in vLLM.
vLLM Qwen3.5 Optimization
Overview
This skill covers Qwen3.5 dense / MoE / GDN runtime, MTP, FP8 and NVFP4 quantization, LoRA, and Eagle3 in vLLM.
Evidence snapshot:
- vLLM mainline checked around
0f7be0f2f76814f80f9091220a5fbbb53912ad00 - Support status: supported on current mainline
- Canonical PR notes:
references/pr-history.md - History mirrors:
model-pr-optimization-history/vllm/qwen35/README.zh.mdandREADME.en.md
Non-Negotiable Evidence Rule
Use skills/model-optimization/model-pr-diff-dossier/SKILL.md as the production bar.
Every PR cited for this family must be based on diff reading, not only PR titles.
Runtime Surfaces
vllm/vllm/model_executor/models/qwen3_5.pyvllm/vllm/model_executor/models/qwen3_5_mtp.py
Current Main Summary
- Qwen3.5 builds on the Qwen3-Next era work but has its own model registration and quantization details.
- The hot spots are GDN fusion, FP8/NVFP4 loading, LoRA target naming, and MoE EP precision.
Key Landed PRs
- #34110
Adding Support for Qwen3.5 Models: Landed the Qwen3.5 runtime family. - #34697
Redo Qwen3.5/Qwen3-Next GDN projector fusion: Reworked an earlier fusion that had to be reverted. - #35289
Fix Qwen3.5 FP8 quantization tuple shard_id weight loading: Closed a concrete FP8 weight-loading failure. - #36658
Add Eagle3 support for Qwen3.5: Enabled the draft-model fast path. - #37975
Extract GatedDeltaNetAttention into shared layer for Qwen3Next and Qwen3.5: Reduced duplicated GDN logic across related families. - #39181
Fix EP precision for Qwen3.5, Qwen3-Next: Patched a serving-precision bug under expert parallelism.
Open Optimization Items
- Re-run PR search before claiming new support beyond the checked mainline commit.
Validation Lanes
- BF16 and FP8 Qwen3.5 serve.
- NVFP4 or MXFP4 quantized startup.
- LoRA and MTP paths.
- Expert-parallel precision regression checks.
Related Skills
BBuf/sglang-humanize-review
development
VerifiedTrustedCommunity
Perform SGLang code review in the style of human maintainers by consulting the full non-agent PR review episode corpus from project start through the latest refresh (June 2026), including inline review threads, top-level PR comments, review submissions, original multilingual text, and multi-round discussions. Use when reviewing SGLang PRs, diffs, patches, or local changes for correctness, tests, performance, GPU/runtime risks, API compatibility, and maintainability.
531SKILL.mdUpdated May 21, 2026
BBuf/model-pr-history-knowledge
documentation
VerifiedTrustedCommunity
Use when an SGLang, vLLM, or TensorRT-LLM serving/model optimization task needs prior model-family PR evidence. Query and read the PR-driven history docs under model-pr-optimization-history before choosing source paths, fast paths, kernel/fusion ideas, regression risks, or validation lanes.
531SKILL.mdUpdated May 17, 2026
BBuf/vllm-sota-humanize-loop
development
VerifiedTrustedCommunity
Run an autonomous Humanize-governed vLLM SOTA performance loop for one LLM model: first perform the fixed fair vLLM/SGLang/TensorRT-LLM deployment search and benchmark, then start one RLCR loop that repeatedly decides the gap, profiles the current bottleneck, runs layer/kernel pipeline analysis, patches vLLM code, optionally uses ncu-report-skill for kernel evidence, and revalidates until vLLM matches or beats the best observed framework under the same workload and SLA.
423SKILL.mdUpdated May 27, 2026
BBuf/llm-pipeline-analysis
devops
VerifiedTrustedCommunity
Inspect LLM torch profiler traces at forward-pass, layer, and kernel level. Use when you need layer timings, anchor-kernel boundaries, representative kernel flows, or Perfetto time ranges.
423SKILL.mdUpdated May 21, 2026