Launch an SGLang server, run a quick accuracy check, and capture an execution trace for performance inspection. Return the trace location and generated files, ready to open in Perfetto or Chrome tracing.
中文简介正在准备中。当前显示英语简介。
Build and debug sglang tests comparing token probabilities during prompt processing and step-by-step generation. Separate numerical differences from reused-state bugs, choose test thresholds, and locate the first operation that diverges.
中文简介正在准备中。当前显示英语简介。
Quantize diffusion transformers with NVIDIA ModelOpt, using FP8 or NVFP4 lower-precision numbers, and prepare the saved weights for SGLang Diffusion. Follow a workflow for loading, checking output quality, and comparing speed under matching conditions.
中文简介正在准备中。当前显示英语简介。
Investigate recurring SGLang automated test failures by comparing logs, commit history, and GPU environments. Narrow down code changes or machine-specific problems, optionally reproduce failures remotely, and summarize the evidence in a structured report.
中文简介正在准备中。当前显示英语简介。
Investigate SGLang serving failures by collecting server state, preserving failing requests, and replaying them on a clean target. Summarize evidence and likely causes, then choose the next debugging tool for crashes, stalls, or performance regressions.
中文简介正在准备中。当前显示英语简介。
Review SGLang Cookbook pull requests for configuration errors, documentation mismatches, benchmark issues, and broken links. Get findings for each file and an overall recommendation on whether to approve, request changes, or block the contribution.
中文简介正在准备中。当前显示英语简介。
Guide test design for SGLang's scripted runtime by deciding when a test harness API adds real value. Favor direct reads, meaningful outcome checks, and the engine's actual execution loop over unnecessary wrappers.
中文简介正在准备中。当前显示英语简介。
Prepare batches of merged SGLang pull requests for a release branch. It checks conflicts and previous attempts, starts the cherry-pick workflow, monitors completion, and reports created pull requests or failure reasons.
中文简介正在准备中。当前显示英语简介。
Attribute GPU time in one sglang-omni stage to specific lines of `sglang_omni/` python source. Drives the `llm-torch-profiler-analysis` backend with a mapping/formal trace pair plus steady-state gates, and carries the rules for reading its three tables on an omni stage. Use when a stage is slower than expected and the question is which python code is the bottleneck, not which kernel is hot.
中文简介正在准备中。当前显示原文说明。
Generate a bounded task plan (and the agent prompt for it) that runs the 5-layer profiling methodology in .claude/skills/model-profiling/METHODOLOGY.md against one model, stops for human confirmation before any GPU work starts, then delegates the actual run to a background agent, verifies its completion claim, and routes findings into a gitignored .profiling-runs/<model>/ directory — durably tracked as a sub-issue under the tracking issue named in SKILL.md's "Result tracking" section — plus the methodology doc.
中文简介正在准备中。当前显示原文说明。
Run all reference benchmarks under benchmarks/eval/ and refresh the reference-table cells in benchmark_*.py. Auto-detects host hardware (H200/H100/H800/...). For each row's matching hw entry: replace cells in place. For new hardware or a new workload tag: append a new row to the section's table. Local-Pipeline-Result tables are skipped. One slash command runs everything end-to-end and auto-commits.
中文简介正在准备中。当前显示原文说明。
Use when you need to CAPTURE / PRODUCE XProf (jax.profiler) profiling artifacts from sgl-jax (SGLang on JAX/TPU) for later performance analysis: driving a live server (/start_profile, stage-separated prefill/decode), offline bench_one_batch --profile, or kernel microbench. This skill produces the trace files; it does not analyze them.
中文简介正在准备中。当前显示原文说明。
Use when evaluating a newly adapted model's accuracy/eval score. Runs evalscope eval (or the test/srt/run_eval.py fallback) against an OpenAI-compatible endpoint on datasets like gsm8k / aime, producing per-dataset reports and a score. Accuracy only — throughput/latency benchmarks belong to model-speed-benchmark. Triggers on accuracy, eval, evalscope, gsm8k, aime, run_eval — not throughput.
中文简介正在准备中。当前显示原文说明。
Use when running throughput/latency (serving speed) benchmarks against an already-running model server. Uses sgl_jax.bench_serving to sweep batch-size/concurrency and measure throughput and latency, producing summary.csv and per-point bs_<N>/result.jsonl. Speed only — accuracy/eval belongs to model-accuracy-benchmark. Triggers on throughput, latency, bench_serving, batch-size sweep, tok/s, TTFT, TPOT — not accuracy/eval.
中文简介正在准备中。当前显示原文说明。
RBG controller stress test - deploy kwok in existing cluster, run create/update/delete load tests with real controller pod, collect pprof profiling and logs, generate analysis report with recommendations
中文简介正在准备中。当前显示原文说明。
Implements Manus-style file-based planning to organize and track progress on complex tasks. Creates task_plan.md, findings.md, and progress.md. Use when asked to plan out, break down, or organize a multi-step project, research task, or any work requiring >5 tool calls. Supports automatic session recovery after /clear.
中文简介正在准备中。当前显示原文说明。
Step-by-step tutorial for adding a new lightweight SYCL JIT kernel for Intel XPU to sgl-kernel-xpu's python/sgl_kernel/jit module (SYCL source + Python wrapper + tests + benchmark)
中文简介正在准备中。当前显示原文说明。
Add a new benchmark to sgl-eval by vendoring its NeMo-Skills dataset module and registering it in `_TABLE`. Use when the user asks to "add <benchmark>" / "support <benchmark>" / "register <benchmark>" inside the sgl-eval repo.
中文简介正在准备中。当前显示原文说明。
Upgrade the vendored NeMo-Skills slice to a newer upstream commit. Use when the user asks to "bump vendored", "sync NeMo-Skills", "update vendored sha", "upgrade NS", or wants to pull a specific upstream fix. Inside the sgl-eval repo.
中文简介正在准备中。当前显示原文说明。
Audit whether all score-deciding logic in sgl-eval is vendored from NeMo-Skills, or whether some has crept into SE code. Use when the user asks "are we vendoring enough", "review vendor coverage", "audit vendoring", "is our vendoring complete", or before a release.
中文简介正在准备中。当前显示原文说明。