{"slug":"mkurman-vllm","source_name":"mkurman/vllm","name":"Mkurman/Vllm","description":"Fast LLM inference engine. PagedAttention, continuous batching, tensor parallelism, speculative decoding, and prefix caching. OpenAI-compatible API server. Supports Llama, Mistral, Qwen, DeepSeek, and hundreds of models.","version":1,"lift":{"pass_rate_delta_pts":4.35,"pass_rate_pct":95.7,"total_cases":23,"passed_cases":22,"tokens_delta_pct":2.8,"turns_delta_pct":0,"verdict":"mixed","benchmark_model":"gemini-3.6-flash","grading_method":"judged","completed_at":"2026-08-21T17:31:06.278281+00:00"},"skill_score":0.9565,"benchmark_models":[{"model":"gemini-3.6-flash","headline":true,"delta_pts":4.35,"with_pass_pct":95.7,"without_pass_pct":91.3,"tokens_delta_pct":2.8,"turns_delta_pct":0,"total_cases":23,"cases_aggregated":23,"verdict":"mixed","never_hurt":false,"completed_at":"2026-08-21T17:31:06.278281+00:00","run_id":"e68ed8c5-a83b-479f-ab8a-d2d94f04857f","version_number":1,"is_latest_version":true,"gate":null}],"trust":{"skill_safety":"passed","safety_status":"clean","intent_verdict":"safe","content_status":"clean","indexable":true},"license":"MIT","install_count":0,"manifest_hash":"6332400e6130c387bb2b6282aadbc0c64be5aeaad5acc9a1c688e048cfdf8728","raw_url":"https://app.decimal.ai/s/mkurman-vllm/SKILL.md","scorecard_url":"https://app.decimal.ai/skills/mkurman-vllm"}