Designs effective prompts for LLM applications, covering structure, context, output format and evaluation.
davila7/claude-code-templates
A collection of skills that helps Claude Code and other coding agents handle AI engineering, product, marketing, data and agent-management tasks.
- 200 skills
- 33.1k stars
The 5 most installed
Ordered by installs on skills.sh- prompt-engineer1.5k installs
- marketing-strategy-pmm1.4k installs
Helps with product positioning, go-to-market plans, launch playbooks, and competitive battlecards.
- google-analytics1.2k installs
Analyzes Google Analytics data to explain traffic, conversions and user behavior, then suggests improvements.
- jira1.2k installs
Creates, views and updates Jira issues and checks sprint status from the conversation.
- ceo-advisor1.1k installs
Guides executive decisions with strategy analysis, financial modeling and board and investor playbooks.
The other 195 skills
- conversation-memory1.1k installs
Adds short-term, long-term and entity-based memory to LLM conversations.
- seo-optimizer1.1k installs
Improves search rankings through keyword research, on-page SEO, meta tags, and technical fixes.
- agile-product-owner988 installs
Writes user stories, plans sprints, manages backlogs and tracks velocity for agile teams.
- copywriting777 installs
Writes or rewrites marketing copy for homepages, landing pages, pricing and product pages.
- marketing-ideas769 installs
Suggests 140 proven marketing tactics organized by category for growing a software product.
- agent-memory-systems728 installs
Explains how to design agent memory, covering short-term context, long-term stores and retrieval strategies.
- content-creator726 installs
Writes SEO-friendly marketing content in a consistent brand voice, with calendars and social templates.
- context7-auto-research724 installs
Fetches the latest library and framework documentation through the Context7 API.
- claude-code-guide699 installs
Explains how to configure and use Claude Code, with prompting strategies and debugging tips.
- ai-agents-architect689 installs
Designs autonomous AI agents with tool use, memory, planning and multi-agent coordination.
- agent-evaluation637 installs
Tests and benchmarks AI agents for capability, reliability and production monitoring.
- prompt-engineering605 installs
Teaches prompt engineering patterns and best practices, and helps debug agent behavior.
- cto-advisor595 installs
Advises on technical debt, team scaling, architecture decisions and engineering metrics.
- context-window-management590 installs
Suggests ways to summarize, trim and route context so long conversations stay within token limits.
- product-manager-toolkit585 installs
Supports product managers with RICE prioritization, interview analysis, PRD templates and go-to-market plans.
- agent-memory-mcp575 installs
Gives agents a persistent, searchable memory for architecture notes, patterns and decisions.
- agent-manager-skill568 installs
Starts, stops, monitors and schedules multiple local command-line agents using tmux sessions.
- seo-audit559 installs
Audits a site for technical and on-page SEO problems and explains why pages aren't ranking.
- product-strategist558 installs
Helps Heads of Product with OKR cascades, market analysis, vision setting and team scaling.
- seo-fundamentals558 installs
Explains SEO basics, E-E-A-T, Core Web Vitals, and how Google's ranking principles work.
- agent-tool-builder537 installs
Guides designing agent tools, from JSON schemas and descriptions to validation and error handling.
- content-research-writer533 installs
Researches topics, adds citations and improves outlines while helping you write content.
- marketing-demand-acquisition525 installs
Plans multi-channel demand generation, paid media, SEO, and partnerships, with a CAC calculator and playbooks.
- autonomous-agents492 installs
Covers agent loops, goal decomposition and reflection patterns for building reliable autonomous agents.
- research-engineer478 installs
Reviews and implements technical work with strict scientific rigor and formal verification.
- app-builder476 installs
Builds full-stack applications from a natural language request, choosing the stack and coordinating agents.
- parallel-agents466 installs
Orchestrates multiple agents to work on independent tasks or gather several perspectives at once.
- computer-use-agents456 installs
Builds AI agents that view screens, move cursors, click and type like a person.
- langgraph447 installs
Builds stateful multi-actor AI agents with LangGraph graphs, checkpoints and human-in-the-loop steps.
- autonomous-agent-patterns446 installs
Explains design patterns for autonomous coding agents, including tools, permissions and browser automation.
- marketing-psychology443 installs
Applies 70+ mental models and behavioral science principles to marketing decisions and messaging.
- prompt-caching440 installs
Explains caching strategies for LLM prompts, including Anthropic prompt caching and cache-augmented generation.
- prompt-library425 installs
Provides ready-to-use prompt templates for coding, writing, analysis and creative tasks.
- pricing-strategy419 installs
Helps decide pricing, packaging and monetization, covering tiers, freemium, value metrics and willingness to pay.
- ml-paper-writing417 installs
Drafts publication-ready ML papers for major conferences, with LaTeX templates and citation checks.
- app-store-optimization404 installs
Researches, optimizes and tracks mobile app listings on the Apple App Store and Google Play.
- micro-saas-launcher401 installs
Helps validate an idea and launch a small SaaS product, covering MVP, pricing, and launch strategy.
- ai-product399 installs
Guides building AI features with LLM integration, retrieval, prompt design, AI UX and cost control.
- perplexity395 installs
Searches the web and researches topics using Perplexity AI for general questions and latest information.
- qa-test-planner395 installs
Writes test plans, manual test cases, regression suites and bug reports, with Figma design validation.
- subagent-driven-development395 installs
Executes implementation plans by splitting independent tasks across subagents in the current session.
- rag-engineer392 installs
Guides building retrieval-augmented generation systems, covering embeddings, vector databases, chunking and retrieval.
- gemini388 installs
Runs the Gemini CLI to review code or plans and process very large contexts.
- lead-research-assistant386 installs
Finds target companies that fit your business and suggests ways to reach them.
- ai-wrapper-product379 installs
Helps build focused products around AI APIs, covering prompts, cost, rate limits and positioning.
- dispatching-parallel-agents379 installs
Splits two or more independent tasks across parallel agents that need no shared state.
- competitive-ads-extractor373 installs
Pulls competitors' ads from ad libraries and analyzes the messaging and creative that works.
- page-cro373 installs
Improves conversion rates on homepages, landing pages, pricing pages, feature pages, and blog posts.
- schema-markup373 installs
Adds and fixes structured data markup so pages can show rich snippets in search results.
- copy-editing366 installs
Edits and polishes existing marketing copy through several focused review passes.
- behavioral-modes364 installs
Switches the agent between working modes such as brainstorm, implement, debug, review and teach.
- analytics-tracking361 installs
Sets up, improves or audits analytics tracking such as GA4 events, conversions and UTM parameters.
- ab-test-setup356 installs
Plans and designs A/B tests and experiments, including variants, hypotheses and measurement.
- programmatic-seo355 installs
Plans SEO pages generated at scale from templates and data, such as location or comparison pages.
- voice-ai-development352 installs
Builds voice AI apps with OpenAI Realtime, Vapi, Deepgram, ElevenLabs, and LiveKit.
- voice-agents349 installs
Builds low-latency voice agents that handle natural conversation, interruptions, and background noise.
- email-systems348 installs
Plans transactional email, marketing automation and deliverability infrastructure that reaches inboxes.
- paid-ads345 installs
Plans and optimizes paid ad campaigns on Google, Meta, LinkedIn, and X, including ad copy and targeting.
- launch-strategy342 installs
Plans product launches and feature announcements with phased rollouts and channel strategy.
- crewai337 installs
Helps build role-based multi-agent teams with CrewAI, including tasks, processes and memory.
- datadog-cli334 installs
Searches logs, queries metrics and traces requests in Datadog from the command line.
- email-sequence332 installs
Creates and optimizes drip campaigns, welcome sequences and lifecycle email flows.
- llm-app-patterns326 installs
Provides production patterns for RAG pipelines, agent architectures and LLM monitoring.
- competitor-alternatives309 installs
Writes competitor comparison and alternative pages for SEO and sales.
- agent-management296 installs
Creates, lists, restarts and manages AI agents through the AI Maestro command line.
- langfuse291 installs
Traces, evaluates and manages prompts for LLM apps using the Langfuse platform.
- signup-flow-cro288 installs
Optimizes signup, registration, and trial activation flows to reduce dropoff and raise conversions.
- paywall-upgrade-cro286 installs
Creates and optimizes in-app paywalls, upgrade screens, upsell modals, and feature gates to convert free users to paid.
- loki-mode285 installs
Orchestrates many specialized agents to take a product spec through build, review, and deployment.
- onboarding-cro283 installs
Optimizes post-signup onboarding, user activation, and first-run experience to improve time-to-value.
- free-tool-strategy276 installs
Plans free calculators or generators you can build to attract leads and search traffic.
- gepetto234 installs
Produces detailed, sectioned implementation plans through research, interviews and multi-model review.
- popup-cro218 installs
Creates and optimizes popups, modals, slide-ins, and banners to capture leads and improve conversions.
- form-cro211 installs
Suggests changes to lead, contact, demo, survey and checkout forms to raise completion rates.
- referral-program207 installs
Designs and optimizes referral, affiliate, and word-of-mouth programs with incentive structures.
- planning190 installs
Creates and maintains markdown plan files to track progress on long, multi-step tasks.
- memory-search139 installs
Searches past conversations and stored memory to recall earlier discussions, decisions, and context.
- agent-messaging108 installs
Sends and receives signed messages between AI agents using the Agent Messaging Protocol.
- docs-search96 installs
Searches generated codebase documentation for function signatures, APIs and class definitions.
- openai-docs88 installs
Answers questions about building with OpenAI products and APIs using up-to-date official documentation with citations.
- graph-query65 installs
Queries a code graph database to show what depends on a component before you change it.
- data-scientist38 installs
Performs advanced analytics, predictive modeling and statistical analysis on your data.
- linkedin-profile-optimizer36 installs
Improves a LinkedIn headline, About section and profile so recruiters and searches find it.
- x-twitter-scraper35 installs
Searches tweets, looks up profiles, extracts followers, and monitors accounts through the X API.
- infrastructure-modal27 installs
Runs ML workloads on serverless GPUs and deploys models as APIs without managing infrastructure.
- owasp-security27 installs
Reviews code and systems against OWASP security standards for web, mobile, API, Kubernetes, and AI agents.
- data-engineer26 installs
Builds data pipelines, warehouses and streaming architectures with Spark, dbt and Airflow.
- deep-research26 installs
Runs autonomous research that plans, searches, reads and writes a comprehensive report.
- ml-engineer24 installs
Builds production machine learning systems with PyTorch and TensorFlow, including serving and monitoring.
- academic-cv-builder22 installs
Formats academic CVs listing publications, grants, teaching and research experience for faculty and postdoc jobs.
- cover-letter-generator18 installs
Writes a personalized cover letter from your resume and a job description.
- multimodal-whisper17 installs
Transcribes and translates speech in 99 languages and identifies the spoken language.
- prompt-engineering-patterns17 installs
Teaches advanced prompt engineering techniques to improve LLM performance, reliability and control.
- deep-research-notebooklm15 installs
Runs multi-source research with NotebookLM and delivers briefs, slides, audio or reports.
- product-decision-agent14 installs
Acts as a senior Chinese internet product lead, diagnosing real problems and recommending the next step.
- seoagent14 installs
Runs a persistent SEO workflow that audits sites, plans keywords, and drafts articles across sessions.
- job-description-analyzer13 installs
Scores how well you match a job posting, points out gaps and suggests an application strategy.
- interview-prep-generator12 installs
Turns a resume into STAR stories, practice questions and talking points for interviews.
- llm-ops12 installs
Guides building production AI systems with RAG, embeddings, vector databases, fine-tuning, and quality evaluations.
- multimodal-audiocraft12 installs
Generates music and sound effects from text descriptions using AudioCraft's MusicGen and AudioGen.
- creative-portfolio-resume11 installs
Builds a distinctive resume for creative roles that still passes automated screening.
- portfolio-case-study-writer11 installs
Turns resume bullets into detailed portfolio case studies with context, action, and outcome.
- llm-evaluation10 installs
Sets up evaluation strategies for LLM apps, from automated metrics to human review and A/B tests.
- career-changer-translator9 installs
Translates your skills between industries and identifies transferable strengths for a career change.
- gemini-api-agent-platform9 installs
Guides building enterprise AI apps with the Gemini API and Google Gen AI SDK.
- pydantic-ai8 installs
Builds production AI agents with PydanticAI, including type-safe tools, structured outputs and multi-model support.
- executive-resume-writer6 installs
Writes C-suite and VP resumes that highlight strategic leadership and board-level impact.
- reference-list-builder5 installs
Formats a professional reference list and prepares referee details for job applications.
- offer-comparison-analyzer4 installs
Compares multiple job offers side by side, including salary, equity, benefits, and total compensation.
- ai-release-hunter3 installs
Collects a daily digest of AI releases from major labs and news sources, then emails it.
- fine-tuning-unsloth3 installs
Guides faster, lower-memory fine-tuning with Unsloth and LoRA optimization.
- mlops-mlflow3 installs
Tracks ML experiments, versions models in a registry, and deploys them to production with MLflow.
- agents-langchain2 installs
Builds LLM apps with agents, chains and retrieval using LangChain and many model providers.
- agents-llamaindex2 installs
Builds document question-answering and retrieval pipelines with LlamaIndex connectors and indexes.
- inference-serving-vllm2 installs
Serves LLMs with high throughput using vLLM batching and quantization options.
- mlops-weights-and-biases2 installs
Logs ML experiments, tunes hyperparameters with sweeps, and manages a model registry in W&B.
- prompt-engineering-instructor2 installs
Extracts structured data from LLM responses with Pydantic validation and automatic retries using Instructor.
- inference-serving-llama-cpp1 installs
Runs LLM inference on CPU, Apple Silicon and consumer GPUs without NVIDIA hardware.
- inference-serving-sglang1 installs
Serves LLMs with fast structured output and prefix caching for JSON and agent workflows.
- inference-serving-tensorrt-llm1 installs
Optimizes LLM inference on NVIDIA GPUs with TensorRT for high throughput and low latency.
- infrastructure-lambda-labs1 installs
Sets up dedicated GPU cloud instances for ML training and inference with SSH access.
- infrastructure-skypilot1 installs
Runs training and batch jobs across multiple clouds while optimizing GPU costs.
- mechanistic-interpretability-nnsight1 installs
Guides running interpretability experiments on large models, including remote execution without a local GPU.
- mechanistic-interpretability-pyvene1 installs
Guides causal interventions on PyTorch models, such as activation patching and causal tracing.
- mechanistic-interpretability-saelens1 installs
Guides training and analyzing sparse autoencoders to break model activations into interpretable features.
Guides inspecting and manipulating transformer internals to study attention patterns and model algorithms.
- mlops-tensorboard1 installs
Visualizes training metrics, compares experiments, and profiles model performance with TensorBoard.
- model-architecture-litgpt1 installs
Implements and fine-tunes LLMs with LitGPT across 20+ pretrained architectures using LoRA or QLoRA.
- tokenization-sentencepiece1 installs
Trains language-independent tokenizers with SentencePiece, including multilingual and CJK text.
Builds and deploys continuous autonomous agents and visual multi-step workflow automations.
Builds teams of specialized AI agents that collaborate on complex tasks with roles and memory.
Lists an AI product on free directories and launch sites one at a time, recording each submission.
Applies Anthropic's official brand colors and typography to documents and designs.
Applies Anthropic's official brand colors and typography to documents and designs.
Cleans and deduplicates large training datasets on GPUs, with quality filters and PII redaction.
Runs scalable batch data processing across CPUs and GPUs with Ray Data.
Adds distributed training to a PyTorch script with a few lines and one launch command.
Guides DeepSpeed setup for ZeRO optimization, pipeline parallelism and mixed precision training.
Trains very large language models with NVIDIA Megatron-Core parallelism strategies.
Guides fully sharded data parallel training in PyTorch, including sharding and offloading.
Sets up PyTorch Lightning training loops with automatic distributed training and callbacks.
Orchestrates distributed training and hyperparameter tuning across many machines with Ray Train.
Compresses large models into smaller student models while keeping most of their performance.
Extends transformer context windows with RoPE, YaRN and position interpolation techniques.
Merges fine-tuned models with mergekit to combine capabilities without retraining.
Shrinks and speeds up language models with pruning methods like Wanda and SparseGPT.
Trains mixture-of-experts models with DeepSpeed or HuggingFace to cut compute costs.
Speeds up model inference with speculative decoding, Medusa heads and lookahead decoding.
Benchmarks code generation models on HumanEval, MBPP and other coding tests.
Benchmarks language models across 60+ academic tests like MMLU and GSM8K.
Runs large-scale model evaluations across 100+ benchmarks on Docker, Slurm or cloud.
Guides fine-tuning with Axolotl using YAML configs, LoRA and preference methods.
Guides fine-tuning with LLaMA-Factory, including a no-code WebUI and quantized LoRA.
Fine-tunes large models with LoRA and other parameter-efficient methods on limited GPU memory.
Finds open US healthcare jobs and the hourly pay employers post for them.
Explains and applies the Mamba state-space architecture for fast, long-sequence language models.
Walks through a compact educational GPT implementation for learning how transformers work from scratch.
Explains and applies the RWKV hybrid architecture for linear-time inference with long context.
Guides distributed LLM pretraining with torchtitan across many GPUs using 4D parallelism.
Handles image captioning, visual question answering, and image-text retrieval with BLIP-2.
Matches images and text for zero-shot classification, image search, and content moderation with CLIP.
Builds vision-language chatbots that answer questions about images and hold multi-turn image conversations.
Segments any object in an image using points, boxes, masks, or automatic mask generation.
Generates images from text prompts and edits them with image-to-image and inpainting pipelines.
Traces, evaluates, and monitors LLM applications to help debug and test model outputs.
Traces, evaluates, and monitors LLM applications with an open-source observability platform.
Compresses large language models to 4-bit with activation-aware quantization for faster inference on limited GPUs.
Quantizes LLMs to 8-bit or 4-bit to cut GPU memory use by 50-75% with little accuracy loss.
Speeds up transformer attention and reduces memory use for long sequences during training and inference.
Converts and quantizes models to GGUF format for efficient CPU and GPU inference on consumer hardware.
Applies 4-bit post-training quantization to large models for big memory savings and faster inference.
Quantizes models to 4, 3, or 2-bit precision without needing calibration data.
Guides GRPO and reinforcement learning fine-tuning with TRL for reasoning and task-specific models.
Guides reinforcement learning training of large MoE models with miles, including FP8/INT4 and speculative RL.
Sets up high-performance RLHF training with OpenRLHF for PPO, GRPO, RLOO and DPO on large models.
Aligns language models with SimPO preference optimization, a simpler reference-free alternative to DPO.
Guides reinforcement learning post-training of LLMs with slime, built on Megatron and SGLang.
Guides PyTorch-native agentic RL training with Meta's torchforge library and its clean abstractions.
Fine-tunes LLMs with TRL for SFT, DPO, PPO/GRPO and reward model training.
Guides large-scale reinforcement learning post-training of LLMs with the verl framework.
Builds modular RAG systems and agents with DSPy and optimizes prompts automatically.
Constrains LLM output with regex and grammars to guarantee valid JSON, XML or code.
Guarantees valid JSON, XML or code output and type-safe results with the Outlines library.
Stores embeddings and metadata in Chroma and runs vector or full-text search for RAG apps.
Builds fast similarity search over large vector collections using Facebook's FAISS library.
Guides you through building retrieval-augmented generation with chunking, embeddings, and vector stores.
Sets up Pinecone as a managed vector database for production search and recommendations.
Builds fast vector search and hybrid filtering with the Qdrant vector database.
Generates text and image embeddings with Sentence Transformers for search and clustering.
Applies Anthropic's constitutional AI method to train models that produce fewer harmful outputs.
Filters model inputs and outputs for safety categories using Meta's LlamaGuard moderation model.
Adds runtime safety rails to LLM apps with NVIDIA NeMo Guardrails, including jailbreak and PII checks.
Builds fast custom tokenizers with Hugging Face Tokenizers for research and production text processing.