from ai-infra-auto-driven-skills218
Guidance and PR-driven validation steps for optimizing and documenting Qwen3.6 support in vLLM; use when auditing or implementing model-specific changes.
Provides an evidence-driven playbook and validation checklist for auditing, debugging, and adding Qwen3.6-specific optimizations to the vLLM codebase. It collects PR history, registry surfaces, and rules for what counts as acceptable production evidence.
Use this skill when an engineer or reviewer needs to: confirm whether vLLM supports Qwen3.6, prepare a PR to add support, or run a targeted audit that cites concrete diffs rather than high-level claims. It's intended for release engineering, model maintainers, and documentation owners.
references/pr-history.md and localized READMEs that capture historical evidence.Best used by agents with code-reading and PR-audit capabilities (e.g., code-review assistants, CI-integrated bots, or Claude Code/Copilot-like tools) that can parse diffs and run repository searches.
This skill has not been reviewed by our automated audit pipeline yet.
