agent-skills
CLAUDE.md5 starsChanged 30 days agoMITView on GitHub
A curated collection of reusable agent skills for software design, refactoring, and implementation work. Skills are activated automatically when a task matches their description — they are not invoked by name. Each SKILL.md must have YAML frontmatter with name and description. The description is the activation trigger — it must be precise enough that a model can decide whether to invoke the skill from it alone.
In this repository11 other files it gives its agents.
- boostSkill
- data-migration-best-practicesSkill
- ddd-best-practicesSkill
- design-patterns-best-practicesSkill
- flowSkill
- fp-best-practicesSkill
- infrastructure-designSkill
- oop-best-practicesSkill
- refactoring-best-practicesSkill
- rest-api-best-practicesSkill
- tdd-best-practicesSkill
The file
We split this CLAUDE.md into the rules it gives an agent
5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 50
Create skills/<skill-name>/SKILL.md with frontmatter and operating instructions.
- Very specific: line 51
Add references/ documents for guidance that should only load on demand.
- Distinctive: line 52
Update README.md with a short entry under Available Skills.
- Very specific: line 53
Keep references/ focused — each file should address one lens or decision type.
- Distinctive: line 57
Write repository prose, headings, comments, and example messages in English.
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Repository Purpose
A curated collection of reusable agent skills for software design, refactoring, and implementation work. Skills are activated automatically when a task matches their description — they are not invoked by name.
Structure
skills/<skill-name>/
SKILL.md # Main operating instructions and frontmatter (name, description, trigger)
references/ # Detailed guidance loaded on demand by the skill
scripts/ # Optional deterministic automationSkill Anatomy
Each SKILL.md must have YAML frontmatter with name and description. The description is the activation trigger — it must be precise enough that a model can decide whether to invoke the skill from it alone.
---
name: my-skill
description: When to invoke this skill — specific task types and contexts.
---
# Skill Title
...Current Skills
| Skill | When to invoke |
boost | Analyze stack, current/desired architecture and delivery lifecycle, prioritize compatible personal skills, and install complete reviewed coverage as symbolic links |
flow | Coordinate multi-step work, delivery readiness, maintenance, migration, and handoff across software and other artifacts |
oop-best-practices | Naming, object boundaries, value objects, cohesion, message-based design |
refactoring-best-practices | Safe incremental changes to existing/legacy code without breaking behavior |
design-patterns-best-practices | Pattern selection (Strategy, State, Factory, Adapter, Decorator, Composite, GoF, PoEAA, Criteria) |
ddd-best-practices | Domain modeling, Bounded Contexts, Aggregates, Domain Events, CQRS, Event Sourcing, Hexagonal Architecture |
infrastructure-design | Event bus, transactions, caching strategies, database views, read model infrastructure |
tdd-best-practices | Red-Green-Refactor cycle, test doubles (mock/stub/fake), outside-in vs inside-out TDD, BDD, test anti-patterns, test pyramid |
fp-best-practices | Pure functions, immutability, function composition/pipe, algebraic data types (Maybe, Either), managing side effects, higher-order functions (map/filter/reduce), currying, FP in JS/TS/Haskell/Elm/Clojure/F#/Elixir |
rest-api-best-practices | REST endpoint design, HTTP method and status code selection, URL naming and versioning, error response format (RFC 9457), pagination, filtering, authentication, rate limiting, CRUD anti-pattern vs task-based APIs |
Adding a New Skill
- Create
skills/<skill-name>/SKILL.mdwith frontmatter and operating instructions. - Add
references/documents for guidance that should only load on demand. - Update
README.mdwith a short entry under Available Skills. - Keep
references/focused — each file should address one lens or decision type.
Writing Guidelines
- Write repository prose, headings, comments, and example messages in English. Render source quotations and chapter titles in English; preserve source URLs, technical identifiers, proper names, and license notices.
descriptionin frontmatter is the model's decision signal — write it as "Use when…" covering concrete task types.SKILL.mdholds workflow steps and day-to-day rules. Keep it scannable.references/files are loaded only when the skill directs the model there — keep them detailed and focused.- Skills cross-link each other via Related Skills sections rather than duplicating guidance.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.