uv-xiao
Publisher on askill
Use when implementing any feature or bugfix, before writing implementation code
Conduct comprehensive literature reviews (systematic/narrative/scoping) across multiple databases, synthesize findings, and produce a well-cited review document. Use when planning and writing literatu...
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
Create and curate new ML domain knowledge skills in this repo. Use when adding a new `knowledge/ML/*` skill, extending the curated ML taxonomy (model-architecture, training, distributed, serving, pape...
Use when starting any conversation or task to establish how to find and apply relevant `uv-*` skills early (without platform-specific assumptions).
Create and maintain the research-project documentation structure (analysis/features/implementation/progress/workloads/spec/evaluation + feats/ impls/ evals/). Use when starting a new research repo, ev...
Download and deeply read an arXiv paper (given an arXiv URL or id), then write a clear human-facing report with strong storytelling and logical reasoning. Use when asked to summarize/review an arXiv p...
Evolve skills safely from real session feedback by persisting structured learnings (`evolution.json`) and stitching an idempotent 'Learned' section into `SKILL.md`. Supports updating both PKB_PATH can...
Plan the slider workflow end-to-end by selecting which repo skills to run (content-prompts, styled-prompts, styled-artifacts) based on the user’s starting input (materials or existing prompts) and req...
Create structured tutorials from repo analysis and hands-on learning artifacts. Use when turning an ML/LLM codebase understanding into teachable material with objectives, diagrams, runnable examples,...
Create publication-quality scientific diagrams via OpenRouter image models with smart iterative refinement and automated quality review. Use when generating figures for papers/reports/slides (architec...
Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs...
Maintain relationships between pkbllm skills so workflows compose cleanly. Use when adding a new skill or changing workflows and you want to (1) decide which skills should be co-invoked, (2) update SK...
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence...
Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1....
RNN+Transformer hybrid with O(n) inference. Linear time, infinite context, no KV cache. Train like GPT (parallel), infer like RNN (sequential). Linux Foundation AI project. Production at Windows, Offi...
Showing 18 of 23