A curated library of 137 production-grade skills for Claude and other AI coding agents.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when writing project instructions for a coding agent (CLAUDE.md, AGENTS.md, or equivalent). Covers what belongs in them, what does not, structure, and keeping them accurate as the project changes.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when an agent needs state that survives a session or a context compaction. Covers what to persist, file-based memory, structuring notes for retrieval, and preventing memory from becoming stale.
nimadorostkar/Claude-Skills-collection
Skill Claude Code
Use when automating agent behavior with lifecycle hooks. Covers hook events, deterministic enforcement of rules the model should not be trusted to remember, and avoiding hooks that make an agent unusable.
nimadorostkar/Claude-Skills-collection
Skill Claude Code
Use when packaging skills, commands, hooks, and MCP servers into a distributable plugin. Covers manifest structure, bundling, versioning, testing, and distribution through a marketplace.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when writing a new skill for an AI agent. Covers scoping, description writing for reliable triggering, progressive disclosure, and the difference between a skill and a prompt.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when reviewing or improving an existing agent skill. Covers triggering accuracy, content quality, redundancy with the base model, and measuring whether the skill actually changes behavior.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when creating reusable slash commands for an agent. Covers when a command beats a skill, argument handling, composing tool calls, and writing commands that are worth typing.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when delegating work to a separate agent with its own context. Covers when delegation pays off, designing a subagent's scope and prompt, parallel fan-out, and the cost of a cold start.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when building an LLM agent that uses tools over multiple steps. Covers tool design, the agent loop, error recovery, termination, human checkpoints, and knowing when an agent is the wrong architecture.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when managing what an LLM sees. Covers context-window budgeting, retrieval and compaction, memory across turns, tool-result pruning, and the failure modes that come from too much context rather than too little.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when considering fine-tuning a model. Covers when fine-tuning beats prompting or RAG, dataset construction, LoRA and full fine-tuning, evaluation, and the failure modes that waste the effort.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when an LLM feature costs too much. Covers prompt caching, context reduction, model routing, batching, output limits, and finding where the tokens actually go.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when measuring the quality of an LLM feature. Covers building an evaluation set, choosing metrics, LLM-as-judge and its pitfalls, regression testing prompts, and evaluating in production.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when integrating an LLM API into an application. Covers streaming, retries and rate limits, timeouts, caching, fallback across providers, and the production concerns that a tutorial integration ignores.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when building a Model Context Protocol server. Covers tool, resource, and prompt design, transport choice, authentication, error handling, and testing an MCP server against a real client.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when building or operating a machine learning pipeline. Covers feature engineering, training reproducibility, train/serve skew, deployment, monitoring for drift, and retraining.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when choosing which LLM to use for a task. Covers matching capability to task, cost and latency trade-offs, routing between models, benchmark skepticism, and evaluating on your own data.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when writing or improving prompts for a language model. Covers instruction structure, examples, reasoning elicitation, output formatting, and systematically diagnosing why a prompt fails.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when building retrieval-augmented generation. Covers chunking, embedding and hybrid search, reranking, grounding and citation, and diagnosing whether a bad answer is a retrieval failure or a generation failure.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when an LLM must return machine-readable data. Covers schema design for models, native structured-output modes, validation and repair, and extraction that survives contact with messy input.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when designing or reviewing an HTTP API. Covers resource modeling, status codes, pagination, idempotency, versioning, error formats, and the contract details that break clients when you get them wrong.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when designing asynchronous job processing. Covers queue selection, idempotency, retry and backoff policy, scheduling, poison messages, and observability for work that happens outside the request.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when adding or debugging a cache. Covers cache placement, invalidation strategies, stampede protection, TTL selection, and the consistency you are trading away.
nimadorostkar/Claude-Skills-collection
Skill Claude CodeCodex
Use when building Django applications. Covers ORM query performance, model design, migrations, Django REST Framework, security defaults, and testing.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: