AI & ML

izzyfresh Bundle

AlphaEvolve expert consultant grounded strictly in the official reference guide. Answers questions about AE suitability, experiment design, evaluator implementation, EVOLVE-BLOCK placement, scoring strategies, concurrency, LLM mixture selection, troubleshooting, API usage, and domain background. Never speculates beyond the guide content. Triggers on: "AlphaEvolve question", "AE best practices", "AE consultant", "is my problem suitable for AE", "how should I design my evaluator", "why is my experiment stuck", "what concurrency should I use", "AE troubleshooting", "ask about AlphaEvolve", "AlphaEvolve help", "how does AlphaEvolve work", "EVOLVE-BLOCK advice", "scoring function advice", "reward hacking", "AE expert".

izzyfresh Skill

Skill: Vertex AI Model Context Protocol (MCP)

josstei Skill

Agent delegation best practices for constructing effective subagent prompts with proper scoping

stacksjs Skill

Use when working with the Stacks ORM — defining models with defineModel(), model relationships (hasOne, hasMany, belongsTo, belongsToMany, morphOne, hasManyThrough), attributes, traits, factories, computed properties, query building, transactions, or the 50+ built-in models. Covers @stacksjs/orm, storage/framework/orm/, and storage/framework/defaults/app/Models/.

stacksjs Skill

Use when working with TypeScript type definitions in a Stacks application — model types, request types, environment variables, event types, billing types, attribute types, or auto-imported globals. Covers storage/framework/types/ and storage/framework/core/types/src/.

stacksjs Skill

Use when working with data models in Stacks — the defineModel() API, model attributes with validation and factories, relationships (hasOne/hasMany/belongsTo/belongsToMany), traits (useAuth, useUuid, useTimestamps, useSearch, useApi, billable, taggable, categorizable, commentable, likeable, observe), computed properties (get/set), model generation, and the 50+ built-in framework models. Covers model definitions and storage/framework/defaults/app/Models/.

stacksjs Skill

Use when working with routing in a Stacks application — defining routes, HTTP methods, route groups, middleware, named routes, URL generation, request enhancement (Laravel-style input/query/file helpers), response helpers, error responses, route model binding, or rate limiting. Covers @stacksjs/router, routes/, and app/Routes.ts.

stacksjs Skill

Use when working with Stacks server actions — creating actions in app/Actions/, auto-generated API actions from the useApi model trait, the 80+ default framework actions (auth, dashboard, commerce, content, deployment, jobs), action request/response handling, or action registration. Covers @stacksjs/actions and storage/framework/defaults/app/Actions/.

stacksjs Skill

Use when adding a new feature end-to-end in a Stacks application — the complete workflow from model definition through migration, action, route, test, and deployment. Covers the recommended order of operations for building features.

hezaohezao Skill

Create, edit, and evaluate agent skills iteratively.

hezaohezao Skill

验证信源可信度的研究过程知识，指导 agent 在涉及事实声明时交叉核验信源

mturac Bundle

Structured self-debugging workflow for AI agent failures using capture, diagnosis, contained recovery, and introspection reports.

skywatch-bsky Bundle

This skill should be used when working with the Ozone moderation API — query patterns, filter combinations, pagination, write tool conventions, and common recipes. Covers both MCP tool access and direct HTTP API access when MCP server is unavailable. Use when querying or writing to Ozone via MCP tools or HTTP. Does not prescribe a workflow, see working-the-queue for queue triage methodology.

skywatch-bsky Skill

Use when writing or modifying Osprey SML rule files from a validated rule specification. Covers model writing, rule writing, effect wiring, and execution graph wiring. Not triggered on general coding tasks.

skywatch-bsky Skill

Use when writing or modifying Osprey SML rule files from a validated rule specification. Covers model writing, rule writing, effect wiring, and execution graph wiring. Not triggered on general coding tasks.

skywatch-bsky Skill

Reference guide for Ozone MCP tools — query patterns, filter combinations, pagination, write tool conventions, and common recipes. Use when working with the Ozone moderation API via MCP tools. Does not prescribe a workflow — see working-the-queue for queue triage methodology.

saddss Bundle

Test-driven development with red-green-refactor loop. Use when user wants to build features or fix bugs using TDD, mentions "red-green-refactor", wants integration tests, asks for test-first development, or writes tests for benchmark harnesses, profiling scripts, or inference automation.

saddss Bundle

Disciplined diagnosis loop for hard bugs and performance regressions. Reproduce → minimise → hypothesise → instrument → fix → regression-test. Use when user says "diagnose this" / "debug this", reports a bug, says something is broken/throwing/failing, describes a performance regression, or debugs vLLM / SGLang / TRT-LLM / lmdeploy failures, benchmark regressions, or flaky inference runs.

saddss Bundle

Build throwaway prototypes to validate a design before production code. Terminal state-machine for routing/cache/overload logic, or UI variants on one route. Use when prototyping, sanity-checking a state model, exploring routing policies, or saying "prototype this" / "let me play with it".

saddss Skill

Research-before-coding workflow. Search the repo, framework APIs, libraries, and existing patterns before writing custom code. Use when adding a feature to a mature framework (vLLM, SGLang, TRT-LLM, benchmark harness), before creating a utility/helper, or when the user asks to "add X functionality".

saddss Bundle

Use before any creative work — adding features, modifying behavior, or extending a mature framework (vLLM, SGLang, TRT-LLM, benchmark harness). Explores intent, constraints, and design; requires user approval before code.

saddss Skill

Performance analysis coordination workflow. Guides profiling delegation, bottleneck classification (compute / memory / launch / communication / sync), and structured report generation. Use when the user asks to analyze performance, profile a workload, check MFU / SOL, diagnose bottlenecks, understand why a vLLM / SGLang / TRT-LLM serve / lmdeploy run is slower than expected, or interpret an existing .nsys-rep against an SLO. Triggers also include "分析性能瓶颈" / "MFU 多少" / "瓶颈是 compute 还是 memory" / "诊断推理慢"

saddss Skill

Write bite-sized implementation plans from an approved spec. Use after brainstorming or grill-with-docs, before touching code on multi-step features in mature codebases (vLLM, SGLang, inference harness, serving stack patches).

saddss Bundle

Grilling session that challenges a plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against project language and documented decisions, or before major inference-stack or benchmark architecture changes.

saddss Skill

Explore a large codebase in parallel by launching multiple explore subagents that each investigate a different area simultaneously. Use when onboarding onto a new project, understanding architecture, investigating cross-cutting concerns, or mapping vLLM / SGLang / TRT-LLM / FlexKV code paths.

saddss Bundle

Analyze host/CPU overhead in LLM inference (vLLM / SGLang / TRT-LLM serve / lmdeploy / TensorRT-LLM) from nsys traces. Phase 1 (Detection): determine whether host overhead is the bottleneck via a binary YES/NO verdict with metric evidence (GPU idle ratio, host prep exposed ratio, per-phase breakdown). Phase 2 (Root Cause): isolate forward step iterations via allreduce kernel patterns, compare NVTX-instrumented host operations across versions, and pinpoint scheduling / request-management regressions. Usable standalone or as a sub-step of perf-analysis. Triggers: host overhead, inter-step gap, scheduling overhead, forward step isolation, nsys iteration analysis, NVTX breakdown, request management overhead, inference loop overhead, between-step gap, GPU idle, host bottleneck detection, host prep exposed, "GPU 利用率上不去", "Python 调度太慢", "vLLM scheduler overhead".

saddss Skill

Launch multiple explore subagents in parallel to investigate architecture, data models, APIs, and deployment; synthesize an onboarding document. Use when first entering an inference framework repo, benchmark harness, or unfamiliar serving stack.

saddss Bundle

Write GPU kernels using NVIDIA CuTe DSL (CUTLASS 4.x Python API) — NOT for Triton, CUDA C++, or conceptual-only questions. Covers element-wise kernels, GEMM patterns, reductions, memory hierarchy (global/shared/register/TMA), MMA tensor core ops, software pipelining, and framework integration. Use when implementing CUTLASS/CuTe kernels for inference GEMM, MoE, or attention in TRT-LLM / custom serving stacks.

saddss Bundle

Nsight Systems (nsys) CLI for system-level timeline profiling. Use when the user wants to run nsys profile, analyze .nsys-rep reports, use nsys stats/analyze/recipe commands, diagnose GPU idle time from timeline traces, profile distributed training with NCCL overlap analysis, or profile an LLM inference server (vLLM / SGLang / TRT-LLM serve / lmdeploy) and its client (e.g. online_replay.py) end-to-end. Triggers also include "用 nsys 抓 profile" / "看 GPU 空闲" / "排查 e2e 延迟" / "分析 NCCL overlap" / "trace 推理服务". NOT for kernel-level metrics like SOL%, occupancy, or roofline (use perf-nsight-compute-analysis for ncu). NOT for writing or generating kernels. NOT for applying optimizations like CUDA Graphs.

saddss Bundle

OpenAI Triton (@triton.jit) kernel development for inference/training ops. NEVER use for CUDA C++, TileIR, or profiling tools (ncu, nsys). Covers fused elementwise, reductions (softmax, LayerNorm, RMSNorm), tiled GEMM with triton.autotune, and flash attention. Workflow: design, write, verify. Use when implementing custom attention, MoE, or KV-cache kernels in vLLM / SGLang stacks. Triggers include "write triton kernel", "Triton GEMM".

saddss Bundle

Profiles and optimizes LLM inference host / CPU overhead (vLLM / SGLang / TRT-LLM serve / lmdeploy / TensorRT-LLM PyExecutor) using line_profiler (with nsys support planned). Runs iterative profile-analyze-optimize-validate rounds. Use when GPU utilization is low, when perf-host-analysis has concluded "host-bound", or when the user asks to reduce host overhead / Python scheduling overhead / sampler latency / inference loop overhead. Triggers also include: "削 host overhead", "降 launch overhead", "优化 scheduler", "GPU 利用率低怎么提".

saddss Bundle

Code instrumentation for timing workloads. Two scenarios: (1) Training loop — inject manual timing to report per-iteration latency, throughput (samples/sec), and data load time. (2) Standalone kernel/op — write CUDA event timing code with warmup, per-iteration statistics, and anti-pattern avoidance. Also covers NVTX annotation for labeling profiler timelines. NOT for: running or analyzing profiler tools (nsys, ncu, Nsight Systems, Nsight Compute), writing kernels (Triton, CuTe, CUDA), applying optimizations (CUDA Graphs, gradient checkpointing, fusion), or interpreting roofline/SOL% metrics. Triggers: "measure throughput", "benchmark this function", "time my training loop", "samples per second", "NVTX annotate", "instrument my dataloader", "data load time", "kernel timing", "how do I time", inference step timing, vLLM forward latency instrumentation.

saddss Bundle

Build operator-level compute templates from model config; estimate FLOPs, MFU, tensor shapes, and TP/EP what-if for serving shapes. Use before benchmarking to sanity-check compute bounds or when writing REPORT.md theoretical sections.

saddss Bundle

Find the maximum sustainable QPS of an LLM inference service that meets a p50 e2e latency SLO using online_replay.py and a binary search. Use for maximum-QPS benchmarks, SLO-based performance tests, and optional serving configuration or feature tuning on local OpenAI-compatible servers.

saddss Bundle

Framework-independent LLM serving benchmark for SGLang, vLLM, TensorRT-LLM. Config-driven search_space over launch flags under shared workload, GPU budget, and latency SLA. Use for cross-framework deployment comparison, cookbook sweeps, or finding best serve command before QPS binary search. Pairs with model-perf-binary-search for SLO-max-QPS tuning after a winner is chosen.

saddss Skill

Review pull requests and code changes for high-performance Python, PyTorch, CUDA, and distributed systems using a correctness-first P0-P4 rubric. Use when the user asks for a strict production-grade review, especially for inference, GPU, concurrency, or performance-sensitive code.