Agent skills

Skills you can use with AI coding agents, indexed from public GitHub repositories.

Latest Most Popular

pytorch Building and training neural networks with PyTorch. Use when implementing deep learning models, training loops, data pipelines, model optimization with torch.compile, distributed training, or deploying PyTorch models.
itsmostafa/llm-engineering-skills 17
lora Parameter-efficient fine-tuning with Low-Rank Adaptation (LoRA). Use when fine-tuning large language models with limited GPU memory, creating task-specific adapters, or when you need to train multiple specialized models from a single base.
itsmostafa/llm-engineering-skills 17
mlx Running and fine-tuning LLMs on Apple Silicon with MLX. Use when working with models locally on Mac, converting Hugging Face models to MLX format, fine-tuning with LoRA/QLoRA on Apple Silicon, or serving models via HTTP API.
itsmostafa/llm-engineering-skills 17
context-engineering Strategies for managing LLM context windows effectively in AI agents. Use when building agents that handle long conversations, multi-step tasks, tool orchestration, or need to maintain coherence across extended interactions.
itsmostafa/llm-engineering-skills 17
prompt-engineering Crafting effective prompts for LLMs. Use when designing prompts, improving output quality, structuring complex instructions, or debugging poor model responses.
itsmostafa/llm-engineering-skills 17
qlora Memory-efficient fine-tuning with 4-bit quantization and LoRA adapters. Use when fine-tuning large models (7B+) on consumer GPUs, when VRAM is limited, or when standard LoRA still exceeds memory. Builds on the lora skill.
itsmostafa/llm-engineering-skills 17
agents Patterns and architectures for building AI agents and workflows with LLMs. Use when designing systems that involve tool use, multi-step reasoning, autonomous decision-making, or orchestration of LLM-driven tasks.
itsmostafa/llm-engineering-skills 17
rlhf Understanding Reinforcement Learning from Human Feedback (RLHF) for aligning language models. Use when learning about preference data, reward modeling, policy optimization, or direct alignment algorithms like DPO.
itsmostafa/llm-engineering-skills 17
transformers Loading and using pretrained models with Hugging Face Transformers. Use when working with pretrained models from the Hub, running inference with Pipeline API, fine-tuning models with Trainer, or handling text, vision, audio, and multimodal tasks.
itsmostafa/llm-engineering-skills 17
skill-customizer Customize existing skills through iterative improvement based on user feedback and preferences. Use when users want to personalize a skill to match their specific workflow, output preferences, domain requirements, or company standards by forking and iteratively refining an existing skill.
nemori-ai/skills 11
mcp-skill-creator Meta-skill for creating workflow-optimized skills from MCP servers. Use when users want to create a custom skill that integrates one or more MCP servers into a specialized workflow. The user provides MCP server configurations and describes their work scenario (workflow, preferences, SOPs), and this skill generates a new skill with optimized scripts following Anthropic's MCP + code execution best practices.
nemori-ai/skills 11
golang-pro Use when building Go applications requiring concurrent programming, microservices architecture, or high-performance systems. Invoke for goroutines, channels, Go generics, gRPC integration.
DeevsDeevs/agent-system 36
datetime Get current date and time in various formats. Use whenever you need the current date, time, timestamps, or formatted datetime values for any purpose (logging, file naming, scheduling, comparisons, etc.)
DeevsDeevs/agent-system 36
97-dev Apply timeless programming wisdom from "97 Things Every Programmer Should Know" when writing, reviewing, or refactoring code. Use for design decisions, code quality checks, professional development guidance, testing strategies, and workflow optimization.
DeevsDeevs/agent-system 36
dev-experts Apply opinionated developer personas for architecture decisions, production debugging, language-specific code review, comprehensive reviewer passes, and test strategy. Use when you need an architect plan, devops investigation, Rust/Python/C++ review, grumpy reviewer audit, or tester-driven test plan. Triggers: architect, devops, rust-dev, python-dev, cpp-dev, reviewer, tester, pre-merge review, refactor for maintainability.
DeevsDeevs/agent-system 36
polars-expertise This skill should be used when the user asks about Polars DataFrame library (Apache Arrow) for Python or Rust. Triggers: "polars expressions", "lazy vs eager", "scan_parquet streaming", "convert pandas to polars", "pyspark to polars", "kdb to polars", "group_by_dynamic", "rolling_mean", "polars window functions", "asof join", "polars GPU", "polars parquet", "LazyFrame". Time series: OHLCV resampling, rolling windows, financial data patterns. Performance: native expressions over map_elements, early projection, categorical types, streaming.
DeevsDeevs/agent-system 36
venue-expert This skill should be used when the user asks about "market microstructure", "exchange mechanics", "order book", "auction", "NBBO", "Reg NMS", "trading venue", "halt", "LULD", "tick size", "maker-taker", "price-time priority", "SIP", "direct feed", "TRF", "wholesaler", "PFOF", "best execution", "trade-through", "ISO", "opening cross", "closing cross", "NOII", "ITCH", "OUCH", or mentions specific exchanges (Nasdaq, NYSE, CME, Binance, SHFE, DCE, CZCE, CFFEX, INE, etc.). For Chinese futures: "CTP", "综合交易平台", "夜盘", "night session", "看穿式监管", "position limits", "持仓限额", queue position in Chinese markets, or Chinese product codes (rb, cu, sc, if, ic, i, j, ta, ma, etc.). Provides hierarchical venue expertise for research and debugging trading systems.
DeevsDeevs/agent-system 36
anti-ai-slop After working on the code, ensure the branch contains only the minimal, idiomatic changes by removing AI-generated slop introduced on this branch.
DeevsDeevs/agent-system 36
bug-hunters Run systematic bug hunting with spec reconstruction, adversarial validation, and confidence scoring. Use when you want to hunt bugs (not fix them), validate correctness, or run logic-first/code-first investigations. Triggers: bug hunt, spec reconstruction, logic-first, code-first, orchestrator, logic-hunter, cpp-hunter, python-hunter.
DeevsDeevs/agent-system 36
chain-system Create, list, and load multi-session chain links that capture work context for later continuation. Use when you want to save current progress, resume a project from a prior session, or inspect chain history. Triggers: chain link, chain load, chain list, save context, resume work, continue later, session handoff.
DeevsDeevs/agent-system 36
cost-status Configure the Claude Code cost/status line plugin that shows session, monthly, total cost, and context usage. Use when a user wants cost tracking, status line setup, or context usage display. Triggers: cost status, status line, show cost, context usage, Claude cost tracking.
DeevsDeevs/agent-system 36
alpha-squad Run multi-lens hypothesis generation across accounting, flows, networks, microstructure, and causal checks. Use when you need market hypotheses with mechanism, counterparty, and decay. Triggers: alpha squad, fundamentalist, vulture, network-architect, book-physicist, causal-detective, hypothesis generation, mechanism first.
DeevsDeevs/agent-system 36
arxiv-search Search arXiv preprint repository for research papers in physics, mathematics, computer science, quantitative biology, finance, and statistics. Use when finding academic papers, preprints, ML research, scientific publications. Triggers: arxiv, preprint, research paper, academic paper, scientific literature.
DeevsDeevs/agent-system 36
mft-research-experts Run research orchestration for data quality, factor geometry, hypothesis validation, and incident forensics. Use when you need SHIP/KILL/ITERATE decisions with strict validation. Triggers: mft-strategist, data-sentinel, factor-geometer, skeptic, forensic-auditor, research pipeline, hypothesis validation, post-mortem.
DeevsDeevs/agent-system 36