All authors
Claude Skills by lenialno
github.com/lenialno13 skills0 installs0 views
- Roofline Capacity RuСпроектировать on-prem LLM-serving под нагрузку и безопасно выкатить: расчёт GPU, VRAM и TP под модель и контекст, concurrency под SLO, число реплик, выбор модели под evals, план canary-выкатки и отката, фиксация версий, перенос весов в air-gap. Use when asked how many GPUs are needed, which model fits the hardware, how many concurrent Claude Code or Qwen Code users a server handles, how to roll out a new model or runtime version safely, or to estimate cost of dedicated inference. Russian-lan...Votes: 0GitHub stars: 2
- Roofline CapacityDesign on-prem LLM serving for a given load and roll it out safely: GPU, VRAM and TP sizing for the model and context, concurrency against the SLO, replica count, model choice driven by evals, canary rollout and rollback plan, version pinning, moving weights into an air-gapped environment. Use when asked how many GPUs are needed, which model fits the hardware, how many concurrent Claude Code or Qwen Code users a server handles, how to roll out a new model or runtime version safely, or to esti...Votes: 0GitHub stars: 2
- Roofline Incident RuRunbook инцидентов LLM-serving на vLLM/SGLang: endpoint не отвечает, CUDA OOM, очередь растёт, TTFT или TPOT p95 выше SLO, preemptions, таймауты клиентов, tool calls сломались после обновления. Use when the server is down or degraded, requests time out, GPU memory is full, latency percentiles spiked, Claude Code or Qwen Code report errors, or to run the triage sequence task health, metrics, gpu, logs. Russian-language version; the English skill has the same name without -ru.Votes: 0GitHub stars: 2
- Roofline IncidentIncident runbook for LLM serving on vLLM/SGLang: endpoint not responding, CUDA OOM, queue growing, TTFT or TPOT p95 above SLO, preemptions, client timeouts, tool calls broken after an upgrade. Use when the server is down or degraded, requests time out, GPU memory is full, latency percentiles spiked, Claude Code or Qwen Code report errors, or to run the triage sequence task health, metrics, gpu, logs.Votes: 0GitHub stars: 2
- Roofline Perf RuУскорить и измерить LLM-инференс на vLLM/SGLang: бюджет VRAM и KV-кеша, continuous batching, prefix/radix cache, квантование (FP8, AWQ, GPTQ, fp8 KV) и его цена для tool calling, спекулятивное декодирование, TTFT/TPOT/p95, методика бенча (task bench, task vram, task metrics). Use when asked why the server is slow, how to raise throughput or concurrency, which quantization to pick, how to size the KV cache, tune max-num-seqs, max-num-batched-tokens or max-model-len, or how to run and read a be...Votes: 0GitHub stars: 2
- Roofline PerfSpeed up and measure LLM inference on vLLM/SGLang: VRAM and KV cache budget, continuous batching, prefix/radix cache, quantization (FP8, AWQ, GPTQ, fp8 KV) and what it costs tool calling, speculative decoding, TTFT/TPOT/p95, benchmark method (task bench, task vram, task metrics). Use when asked why the server is slow, how to raise throughput or concurrency, which quantization to pick, how to size the KV cache, tune max-num-seqs, max-num-batched-tokens or max-model-len, or how to run and read ...Votes: 0GitHub stars: 2
- Roofline RuМаршрутизатор роли Inference Engineer для on-prem LLM-serving (vLLM/SGLang, OpenAI-совместимый endpoint, air-gap). Use when the user asks what to do first with a local LLM server, how to approach a serving task, which skill, task or agent to use, what is inside or outside the inference engineer role, or talks about LLM serving, GPU inference, inference engineering in general terms. Russian-language version; the English skill has the same name without -ru.Votes: 0GitHub stars: 2
- Roofline Runtime RuПоднять vLLM или SGLang on-prem и держать его в порядке: выбор движка, порядок флагов, профили запуска, air-gap, docker, task serve/stop/logs/flags. Use when asked to start, launch, configure, restart or stop a vLLM or SGLang server, choose between vLLM and SGLang, write a launch profile, set tensor parallel, max-model-len, gpu-memory-utilization, or when the server fails to start (OOM at load, port busy, unsupported model, missing tokenizer). Russian-language version; the English skill has t...Votes: 0GitHub stars: 2
- Roofline RuntimeBring up vLLM or SGLang on-prem and keep it healthy: engine choice, flag order, launch profiles, air-gap, docker, task serve/stop/logs/flags. Use when asked to start, launch, configure, restart or stop a vLLM or SGLang server, choose between vLLM and SGLang, write a launch profile, set tensor parallel, max-model-len, gpu-memory-utilization, or when the server fails to start (OOM at load, port busy, unsupported model, missing tokenizer).Votes: 0GitHub stars: 2
- Roofline ToolboxOperations toolbox for a local vLLM or SGLang server with an OpenAI-compatible API. Taskfile tasks and dependency-free scripts for gpu, health, models, metrics (queue, KV cache, prefix-cache hit rate, TTFT/TPOT p95), serve and stop by named profile, fixed SLO benchmark, tool-calling smoke test, VRAM budget from config.json, runtime flag lookup. Use when asked to check, start, stop, benchmark or probe a local LLM server, or when another roofline-* skill says to run a task.Votes: 0GitHub stars: 2
- Roofline Toolcall RuНастроить tool calling и chat template на vLLM/SGLang для Claude Code, Qwen Code и OpenAI-совместимых прокси: выбор --tool-call-parser и --reasoning-parser по семейству модели, проверка шаблона, стриминг tool_calls, task smoke-toolcall. Use when tool calls come back as plain text, arguments are not valid JSON, a client ignores or breaks on tools, the server returns 400 on role tool messages, thinking tags leak into content, or when connecting Claude Code, Qwen Code or a proxy to a local model...Votes: 0GitHub stars: 2
- Roofline ToolcallConfigure tool calling and the chat template on vLLM/SGLang for Claude Code, Qwen Code and OpenAI-compatible proxies: pick --tool-call-parser and --reasoning-parser by model family, verify the template, stream tool_calls, run task smoke-toolcall. Use when tool calls come back as plain text, arguments are not valid JSON, a client ignores or breaks on tools, the server returns 400 on role tool messages, thinking tags leak into content, or when connecting Claude Code, Qwen Code or a proxy to a l...Votes: 0GitHub stars: 2
- RooflineRouter for the Inference Engineer role on on-prem LLM serving (vLLM/SGLang, OpenAI-compatible endpoint, air-gap). Use when the user asks what to do first with a local LLM server, how to approach a serving task, which skill, task or agent to use, what is inside or outside the inference engineer role, or talks about LLM serving, GPU inference, inference engineering in general terms.Votes: 0GitHub stars: 2