Featured

Deploy OpenClaw in 60 seconds — 20% off logoDeploy OpenClaw in 60 seconds — 20% off

Launch OpenClaw on Hostinger in about 60 seconds and keep your agent live 24/7. Our referral link gives you 20% off, no coupon code needed.

Launch on Hostinger
Run your Hermes agent on Hostinger, fully managed logoRun your Hermes agent on Hostinger, fully managed

Launch Hermes on Hostinger in one click, fully managed, no VPS knowledge needed. Use code ZACAARON10 for 10% off.

Launch on Hostinger
Crawl and scrape any site into clean data, 10% off logoCrawl and scrape any site into clean data, 10% off

Firecrawl crawls and scrapes any site into clean markdown for your agent. Get 1,000 free credits, and new users get 10% off their first purchase.

Try Firecrawl free
6,000+ web scrapers for your AI agent, start free logo6,000+ web scrapers for your AI agent, start free

Apify gives your agent live web data: 6,000+ prebuilt scrapers and actors, MCP-ready. Sign up free with $5 in usage credits.

Try Apify free
One API to scrape, enrich, and extract the internet. logoOne API to scrape, enrich, and extract the internet.

Context.dev gives your agents a single API to scrape, enrich, and extract live web data — no proxies, no parsers, no maintenance.

Start building free
SetupClaw: done-for-you OpenClaw for founders & exec teams logoSetupClaw: done-for-you OpenClaw for founders & exec teams

White-glove OpenClaw for founders and exec teams (4–50+ employees): we install, harden, integrate your tools, and maintain it — secured from day one.

Get it set up for you
SEO data APIs for your agent, $1 free credit logoSEO data APIs for your agent, $1 free credit

DataForSEO gives your agent live access to SERP results, keyword data, backlinks, and on-page SEO data through one API. New accounts get a $1 credit, good for up to 20,000 keyword or backlink lookups.

Try DataForSEO free
Reach 48,000+ AI builders

A flat monthly placement in front of developers actively installing AI tools. No lock-in, cancel anytime.

Advertise here

Summary

Skills for UX research, application design, product overviews, personas, scroll storytelling, and design systems.

Install to Claude Code

/plugin install nexus-design@nexus-agents

Run in Claude Code. Add the marketplace first with /plugin marketplace add scanady/nexus-skills if you haven't already.

README.md

> Agentic skills, prompts, and instructions for AI coding platforms — GitHub Copilot, Claude Code, and Codex.

![Skills](skills/)

Table of Contents

Overview

Curated collection of agent skills, reusable prompts, and coding instructions for AI development platforms. Skills are self-contained behavior packages that give AI agents specialized capabilities — from drafting PRDs and writing SOPs, to building MCP servers and running TDD workflows.

Install any skill once and invoke it in your AI agent by name. Skills work with GitHub Copilot, Claude Code, and Codex.

Features & Capabilities

  • 141 production-ready skills organized by domain: marketing, strategy, product, design, tech, and more
  • Multi-platform support — install to GitHub Copilot, Claude Code, or Codex
  • Project or global scope — install skills for a single project or system-wide
  • CLI installer — install, list, and upgrade skills from the command line
  • VS Code agents — pre-built Copilot agent .agent.md configurations
  • Instruction files — path-scoped coding guidelines for consistent AI behavior across teams
  • Reusable prompts — 25+ structured .prompt.md files for common engineering and product workflows

Architecture Overview

Distribution layer (skills/, bin/, src/) The skills/ folder is the source-of-truth skill registry. Each skill is a self-contained folder with a SKILL.md file and optional references/ materials. The CLI (bin/nxa.js) copies skills from this registry into AI agent config directories on the local machine, with implementation code in src/.

VS Code Copilot customization layer (.github/) The .github/ folder contains VS Code Copilot customizations: agent definitions (.github/agents/), path-scoped instruction files (.github/instructions/), reusable prompts (.github/prompts/), and global Copilot instructions (.github/copilot-instructions.md). These files are picked up automatically by VS Code Copilot and are not installed via the CLI.

Key Concepts

Skills

A skill is a SKILL.md file that gives an AI agent a specialized behavior or workflow. Skills use YAML frontmatter to declare their name, description, and metadata. The description is the search trigger — the agent routes to a skill when the user's request matches it.

Skills are installed into an AI agent's config directory (e.g., .agents/skills/ for GitHub Copilot). Once installed, skills are invoked by name in chat.

Agents

Agents (.github/agents/*.agent.md) are VS Code Copilot custom agent modes. Each agent has a role, system prompt, and optional handoffs to other agents.

Instructions

Instructions (.github/instructions/.instructions.md) are scoped coding guidelines automatically injected by VS Code Copilot for matching file paths. For example, java-backend.instructions.md applies to backend//.java files.

Prompts

Prompts (.github/prompts/*.prompt.md) are reusable task templates invoked with # in VS Code Copilot Chat. Examples: nexus-document-readme.prompt.md, nexus-quality-tdd.prompt.md.

Project Structure

{root}/
├── bin/
│   └── nxa.js                  # Published CLI executable
├── src/
│   ├── cli/                    # CLI parsing, help, and command handlers
│   ├── core/                   # Skill discovery, validation, packs, and file utilities
│   ├── sources/                # Local, git, and GitHub skill sources
│   └── audit/                  # Skill overlap analysis engine
├── scripts/
│   ├── build/                  # Plugin and GitHub Pages builders
│   ├── catalog/                # Skill catalog scan/rebuild scripts
│   ├── export/                 # Manual skill zip export scripts
│   └── validation/             # Repository validation scripts
├── skills/                     # Source skill registry
│   ├── agents-copilot-instructions/
│   ├── content-copy-humanizer/
│   ├── engineering-quality-tdd/
│   ├── product-spec-prd-generator/
│   └── ...                     # 141 skills total
├── prompts/                    # Legacy prompt templates
├── docs/                       # Project documentation
├── .github/
│   ├── copilot-instructions.md # Global Copilot instructions
│   ├── agents/                 # Copilot agent definitions
│   ├── instructions/           # Path-scoped instruction files
│   └── prompts/                # Reusable Copilot prompt files
└── package.json

Prerequisites & Requirements

Installation

From a cloned repo

From the project root, install all skills to the current project (default agent: GitHub Copilot):

node bin/nxa.js install

Install a specific skill to a specific agent:

node bin/nxa.js install --skill content-copy-humanizer -a github-copilot

Install to multiple agents at once:

node bin/nxa.js install -a github-copilot

Install globally instead of project-scoped:

node bin/nxa.js install --skill product-spec-prd-generator --global

CLI Reference

| Command | Short | Description | |---------|-------|-------------| | install | | Install skills | | list | | List available skills | | audit-overlap | | Find duplicate and overlapping skills | | --skill <name> | -s | Install a specific skill (repeatable) | | --pack <name> | -P | Install all skills in a named pack (repeatable) | | --agent <name> | -a | Target agent (repeatable, default: agent-skills) | | --global | -g | Install globally (default: project) | | --project | -p | Install to project (explicit) | | --upgrade | -u | Upgrade existing skills (prompts for confirmation) | | --overwrite | -o | Skip confirmation when upgrading | | --source-url <url> | | Clone skills from a git repository instead of the bundled skills | | --source-ref <ref> | | Branch, tag, or commit to clone with --source-url | | --full | -f | List skills with descriptions | | --names | -n | List skill names only | | --count | -c | Print skill count only |

audit-overlap options:

| Option | Short | Description | |--------|-------|-------------| | --threshold <n> | -t | Minimum overlap score to report (default: 0.20) | | --top <n> | | Limit output to top N pairs | | --json-only | | Write JSON report only | | --md-only | | Write Markdown report only | | --output <dir> | -o | Output directory (default: output/) |

Supported agents

| Agent | Global path | Project path | |-------|-------------|--------------| | agent-skills (default) | ~/.agents/skills/ | .agents/skills/ | | github-copilot | ~/.github/skills/ | .github/skills/ | | claude-code | ~/.claude/skills/ | .claude/skills/ | | codex | ~/.codex/skills/ | .agents/skills/ |

Usage

After installation, invoke a skill in your AI agent chat by describing what you want to do. The agent matches your request to the closest skill by description.

For repeatable team installs, pin a branch, tag, or commit in the git package spec or pass --source-ref so every project pulls the same skill set.

Examples:

/skill-architect create a new skill for drafting weekly status updates
/engineering-api-mcp-builder build an MCP server for the GitHub API with issues and PR tools
/product-spec-prd-generator generate a PRD for a mobile expense tracking app
/engineering-quality-tdd implement a user authentication service

Skill names map directly to the folder names in skills/. Use node bin/nxa.js list --full to see all available skills with their descriptions.

Skills Reference

Review the skills catalog viewer and skill catalog detail. Skills are organized by domain prefix.

Adding Skills

Skill structure

skills/
└── your-skill-name/
    ├── SKILL.md          # Required: skill definition with YAML frontmatter
    └── references/       # Optional: reference materials loaded by the skill

A valid SKILL.md requires at minimum:

---
name: your-skill-name
description: Use when [trigger condition] — [what it does]
---

# Skill Title

[Skill instructions here]

Additional Resources

Automated build pipeline

The repo ships an end-to-end automation that keeps the catalog, plugins, and distribution site consistent with skills/. The same scripts run locally and in CI, so a local build matches what gets published.

Local build

npm run validate        # spec checks against every skills/<skill>/SKILL.md
npm run scan            # writes skill-catalog-scan.json (mechanical metadata)
npm run catalog         # rebuilds skill-catalog.json (merges scan + evaluations + packs)
npm run build:plugins   # regenerates .claude-plugin/marketplace.json + per-pack bundles in dist/plugins/
npm run build:site      # assembles GitHub Pages site in dist/site/ (catalog browser, pack manager, per-pack zips)
npm run check:artifacts # fails if committed generated artifacts are stale

npm run build           # all of the above in order
npm run prepr           # build, then verify committed artifacts match generated output

Before opening a pull request, run npm run prepr and commit any changes to

skill-catalog.json, .claude-plugin/marketplace.json, or

.claude-plugin/plugin.json. To install the same check as a local pre-push hook, run npm run hooks:install once in your clone.

Outputs:

| Path | Source of truth | Committed? | |------|------|------| | skill-catalog-scan.json | npm run scan | no | | skill-catalog.json | npm run catalog (merges scan + skill-catalog-evaluations.json) | yes | | .claude-plugin/marketplace.json | npm run build:plugins (from packs/*.json) | yes | | .claude-plugin/plugin.json | npm run build:plugins (umbrella nexus-all) | yes | | dist/plugins/<pack>/ | npm run build:plugins (self-contained pack bundles) | no | | dist/site/ | npm run build:site (GitHub Pages root) | no |

CI workflows

| Workflow | Trigger | Purpose | |---|---|---| | .github/workflows/ci.yml | PR + push to main | Validate skills, rebuild catalog, generate plugins, build site; warns if committed catalog/marketplace drifts from source. | | .github/workflows/release.yml | push to main | release-please opens a release PR with bumped version + changelog from Conventional Commits. On merge, tags a release and attaches per-pack zips, skill-catalog.json, and marketplace.json as release assets. | | .github/workflows/pages.yml | release published (or manual) | Builds dist/site/ and deploys it to GitHub Pages. |

Releases and changelog

Versioning is driven by Conventional Commits. Use the following prefixes so release-please can categorize commits into the CHANGELOG:

| Prefix | Bump | Changelog section | |---|---|---| | feat: | minor | ✨ Features | | fix: | patch | 🐛 Bug Fixes | | perf: | patch | ⚡ Performance | | skill: | patch | 🧠 New / Updated Skills | | pack: | patch | 📦 Pack Changes | | docs: | patch | 📝 Documentation | | build: / ci: | patch | 🛠️ Build / 🤖 CI | | chore: / refactor: / test: / style: | none | hidden | | feat!: or BREAKING CHANGE: footer | major | top of release notes |

release-please maintains a long-lived PR titled chore(main): release X.Y.Z. Merging that PR cuts the release; nothing else needs to be tagged manually.

Adding a pack

1. Create packs/<your-pack>.json with name, description, and a skills array (literal names or prefix-* globs). 2. Run npm run build locally and commit the regenerated catalog and plugin manifests. Do not commit skill-catalog-scan.json. 3. Open a PR. CI verifies everything compiles; on merge, release-please will pick the change up in the next release.

Related plugins

Browse all →