Mattpocock's Skills Collection: Your New Favorite Development Tool
Mattpocock's Skills Collection: Your New Favorite Development Tool
The mattpocock skills repository offers practical advice and tools for developers looking to enhance their skills.
My current favorite skills repository is undoubtedly the collection of skills by mattpocock, which I have recommended multiple times: https://github.com/mattpocock/skills
These skills are truly crafted by real engineers. Most of them consist of just a few simple sentences, but they provide genuine experience and practical development advice. Additionally, they are a key way to learn his programming philosophy. Whether it's /grill-me, /caveman, /zoom-out, or /hand-off — I use them all regularly.
Particularly good is the skill /to-tickets. Imagine: you have already worked through an idea, gone through /grill-me — its skill that helps align requirements between you and the agent before starting work — and now you have a huge implementation plan in front of you. If you hand this plan directly to the agent, the likelihood of getting a mediocre result is very high. The reason is simple: you overload the context by trying to do everything at once. This is where /to-tickets comes in handy. It breaks down the large plan into smaller tasks, marking each with the tag ready for agent.
After this, the agent can work sequentially on each task, clearly understanding what the AC (Acceptance Criteria) are, and is able to verify the result independently (or run a sub-agent for review).
☕️
Why it matters
AnalysisThis repository not only provides useful tools for developers but also helps structure workflows, which can significantly enhance team efficiency.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
Part II. AI Agent Architecture: From Model to Memory and Tools
A deep engineering dive: tokenization, KV-cache, sampling determinism, the training stack, quantization, context engineering, the safety stack, production RAG, ReAct, and multi-agent swarm systems.
15 Claude Cowork Skills Pack: ready-to-use skills for work and automation
A practical set of 15 ready-to-use Agent Skills for Claude Code and Cowork: site automation, contract review, invoice generation, and analytics.
Dynamic Tool Selection
An architectural approach for building scalable AI agents equipped with hundreds of tools. Instead of loading all function descriptions into the context simultaneously, the system employs semantic search or a Router model for dynamic selection of only 3-5 most relevant tools for a specific user query.
Function Calling / Tool Calling
A technical protocol and standard for LLM interaction with external software. Instead of free text, the model returns valid JSON containing the function name and typed arguments according to JSON Schema, enabling the backend to deterministically execute actions in real APIs.