One command now gives a coding agent 131 open-licensed engineering skills from Tech Leads Club, Matt Pocock, Cloudflare, Thomas Ricouard, Jesse Vincent and Anthropic. They cover planning, debugging, code review, architecture, refactors and deploys, and they're packaged as the Dev Workflow plugin in Cuttlefish.MD, HackerNoon's license-checked skills marketplace.
An agent skill is a folder with a SKILL.md file of instructions, plus any scripts or references it needs. The agent loads it only when a task calls for it. The format is an open standard at agentskills.io that 32+ tools support, including Claude Code and Codex CLI.
Every skill below comes from HackerNoon's SKILL.md Library, is MIT- or Apache-licensed and links to its source. They're grouped by author so you can see whose habits you're installing.
Tech Leads Club's agent-skills is the deepest set here. It covers architecture, deploys, Nx monorepos, web performance and a plan-to-implement pipeline (the tlc- skills).
|
Skill |
What it does |
|---|---|
|
AWS architecture design, security review and implementation guidance | |
|
Deploys apps and infrastructure to Cloudflare Workers and Pages | |
|
Finds its way around an unfamiliar codebase | |
|
Behavioral guidelines that cut common LLM coding mistakes | |
|
Finds duplicate business logic across components and suggests consolidation | |
|
Detects misplaced classes and fixes component hierarchy problems | |
|
Maps components and measures their size to decide what to extract first | |
|
Works in Confluence through the Atlassian MCP server | |
|
Optimizes LCP, INP and CLS for page experience and ranking | |
|
Scores coupling between modules on strength, distance and volatility | |
|
Writes Architecture Decision Records for significant choices | |
|
Creates Cursor subagents with isolated context for multi-step work | |
|
Plans step-by-step migrations out of a monolith | |
|
Maps business domains and suggests service boundaries with DDD strategic design | |
|
Groups components into business domains for a service-based architecture | |
|
Designs evolutionary modular-monolith platforms with strategic and tactical DDD | |
|
Generates Excalidraw diagrams from plain-language descriptions | |
|
Addresses review comments on the current branch's PR with the gh CLI | |
|
Debugs and fixes failing GitHub Actions checks on a PR | |
|
Audits AGENTS.md, rules and skills for broken paths and redundant instructions | |
|
Searches, creates and moves Jira issues through the Atlassian MCP server | |
|
Plans legacy modernization, monolith decomposition and microservice consolidation | |
|
Creates, validates and renders Mermaid diagrams to SVG, PNG or ASCII | |
|
Runs a sequenced monolith-to-modular pipeline from inventory to plan | |
|
Builds modular monoliths in NestJS with DDD and clean architecture | |
|
Deploys web projects with the Netlify CLI | |
|
Senior-operator mode: the agent resolves tasks end to end without hand-holding | |
|
Watches Nx Cloud CI and applies self-healing fixes | |
|
Scaffolds projects, libraries and features with Nx generators | |
|
Runs build, test, lint and serve across an Nx workspace | |
|
Configures and optimizes Nx monorepos | |
|
Designs partner, affiliate and integration programs | |
|
Astro-specific optimizations for 95+ Lighthouse scores | |
|
Runs Lighthouse audits, interprets reports and sets performance budgets | |
|
Cuts bundle size and tunes images, caching and lazy loading | |
|
Opinionated Rails conventions: rich models, concerns, CRUD everything, Minitest | |
|
React and Next.js performance guidelines from Vercel Engineering | |
|
React composition patterns that scale | |
|
Builds production React Native and Expo apps | |
|
Deploys to Render by generating render.yaml Blueprints | |
|
Summarizes production errors and health data from Sentry (read-only) | |
|
Shopify reference: Liquid, OS 2.0 themes, GraphQL APIs, Hydrogen, Functions | |
|
Creates subagents with isolated context for complex workflows | |
|
Finds anemic domain models and refactors them into rich ones | |
|
Plays devil's advocate against ideas, plans and proposals | |
|
Reaches a verdict on a decision or tradeoff from several perspectives | |
|
Interviews a rough idea into a verdict and a design document | |
|
Implements a planned ticket by turning it into checks with proofs | |
|
Turns a ticket, PRD or thread into one plan a human approves | |
|
Interviews an idea into a PRD: problem, success, boundaries, journeys | |
|
Spec-driven feature work with EARS acceptance criteria | |
|
Deploys apps and sites to Vercel | |
|
Audits performance, accessibility, SEO and best practices in one pass |
Matt Pocock's skills push the agent to interview you before it builds, then turn the answers into specs, tickets and code.
|
Skill |
What it does |
|---|---|
|
Tells you which skill or flow fits your situation | |
|
Pursues a long-running goal in one session by coordinating subagents | |
|
Hands the conversation to a fresh background agent that picks up the work | |
|
Reviews changes since a commit, branch or tag for standards and correctness | |
|
A diagnosis loop for hard bugs and performance regressions | |
|
Sets up Claude Code hooks that block push, reset --hard and other dangerous git commands | |
|
A relentless interview that sharpens a plan or design | |
|
Grills you about a plan, decision or idea | |
|
Compacts the conversation into a handoff document for another agent | |
|
Implements work from a spec or set of tickets | |
|
Implements the output of to-spec and to-tickets | |
|
Finds deepening opportunities and presents them as a visual HTML report | |
|
Grills you on specs for the workflows you want to build | |
|
Migrates test files from | |
|
Writes a PR description | |
|
Builds a throwaway prototype to answer a design question | |
|
Investigates a question against primary sources and saves the findings in the repo | |
|
Runs a retrospective on a coding session | |
|
Sets up a repo's issue tracker, triage labels and domain docs for these skills | |
|
Adds Husky pre-commit hooks with Prettier, type checks and tests | |
|
Wires dependency-cruiser in so each TypeScript package stays a deep module | |
|
Shared language for designing a software factory and its services | |
|
Turns a decision you can't answer into a questionnaire for someone else | |
|
Turns the conversation into a spec on your issue tracker | |
|
Breaks a plan into tracer-bullet tickets with their blocking dependencies | |
|
Moves issues and outside PRs through triage: categorize, verify, respond | |
|
Makes the agent stop and check before it carries on | |
|
Plans work too big for one session as a map of decision tickets | |
|
Generates a bash wizard that walks a human through steps only they can do |
Cloudflare's official skills teach an agent to build, debug and deploy on its platform.
|
Skill |
What it does |
|---|---|
|
Builds, debugs and reviews apps on the Cloudflare Agents SDK | |
|
Uses Cloudflare's official cf CLI (beta) for the API and Workers projects | |
|
Picks the right Cloudflare products for apps, APIs, agents and storage | |
|
Designs and troubleshoots Cloudflare One Zero Trust and SASE setups | |
|
Plans migrations from VPN, SWG or SASE platforms to Cloudflare One | |
|
Builds and debugs Durable Objects for persistent state | |
|
Builds Workers invoked by Gateway HTTP policies | |
|
Builds and troubleshoots K2 and K2 Streams durable logs | |
|
Builds and deploys Next.js apps on Workers with vinext | |
|
Migrates Sandbox apps to the SDK 1.0 preview | |
|
Builds Sandbox apps on the SDK 1.0 preview | |
|
Builds Sandbox apps on the stable SDK | |
|
Sets up or repairs Turnstile bot verification front and back end | |
|
Diagnoses and fixes Core Web Vitals and Lighthouse performance | |
|
Production best practices for Workers | |
|
Profiles CPU and memory in deployed Workers and Durable Objects | |
|
Runs and troubleshoots Wrangler CLI commands |
Thomas Ricouard's skills lean toward Swift, iOS and macOS, plus parallel multi-agent review and debugging.
|
Skill |
What it does |
|---|---|
|
Writes App Store release notes from every user-facing change since the last tag | |
|
Runs parallel read-only agents to find a bug's root cause | |
|
Works with GitHub through the gh CLI | |
|
Builds, runs and debugs the iOS project on a simulator with XcodeBuildMCP | |
|
Builds and reviews macOS menubar apps that use Tuist and SwiftUI | |
|
Scaffolds and packages SwiftPM macOS apps without an Xcode project | |
|
Plans and runs large refactors with parallel agents | |
|
Reviews past sessions and memory files to recommend the skills worth writing | |
|
Fixes slow renders, re-render thrash and laggy lists in React | |
|
Reviews a diff for reuse, quality and clarity, then simplifies it | |
|
Runs parallel read-only agents over a diff to catch regressions | |
|
Reviews and fixes Swift 6.2+ concurrency | |
|
Builds and reviews SwiftUI features with the iOS 26 Liquid Glass API | |
|
Audits and improves SwiftUI runtime performance | |
|
Refactors SwiftUI views into small subviews with MV-over-MVVM data flow |
Superpowers makes the agent brainstorm, plan, test and verify before it calls anything done.
|
Skill |
What it does |
|---|---|
|
Finds out what you actually want before any feature or component gets built | |
|
Explains why a session went wrong: repeated work, ignored plans, stumbles | |
|
Splits two or more independent tasks across parallel agents | |
|
Works through an implementation plan step by step in the current session | |
|
Decides how to merge finished, tested work | |
|
Checks review feedback for technical merit before acting on it | |
|
Gets work reviewed against requirements before merging | |
|
Hands each task in a plan to a fresh subagent | |
|
Finds the root cause of a bug or failing test before proposing a fix | |
|
Isolates feature work in its own git worktree | |
|
Teaches the agent to find and use skills at the start of every conversation | |
|
Runs the checks before the agent claims anything is fixed or passing | |
|
Writes a step-by-step plan from a spec before touching code |
These come from Anthropic's skills repo, and all four are Apache-2.0.
|
Skill |
What it does |
|---|---|
|
Reference for the Claude API and SDKs: models, streaming, tool use, MCP, caching | |
|
Prompts you to check an answer or draft before you act on it | |
|
Guides you through building high-quality MCP servers | |
|
Creates, improves and measures skills |
/plugin marketplace add hackernoon/cuttlefish-agent-skills
/plugin install dev-workflow@cuttlefish
For Codex CLI, Cursor, Gemini CLI or another agent that supports the standard, clone the repo and copy plugins/dev-workflow/skills/ into your agent's skills folder. Codex reads .agents/skills/ in a project or ~/.agents/skills/ globally.
Each skill keeps its license and an ATTRIBUTION.md naming its author and source commit. Read more under #agent-skills on HackerNoon, browse the full 2,908-skill library at hackernoon.com/skills, and if one of these saves you an hour, star Cuttlefish.MD on GitHub.