🧠 AI Agent Skills Hub — 142+ Production-Grade Skills for Claude Code, Codex & MCP Agents

A single-source AI agent skills aggregator — install, update, and manage 142+ curated, production-grade agent skills for Claude Code, Codex, Antigravity, OpenCode, Cursor, Windsurf, and any MCP or AI agent ecosystem. Skills are auto-synced daily from 12 upstream repositories so you always get the latest versions from one command: npx skills add harshsinghmp/skills.

⚡ Quick Start — Install Skills with `npx skills`

No cloning, no manual configuration. Install any of the 142+ AI agent skills directly into your workspace or globally:

1. Interactive Multi-Select (pick the skills you need)

npx skills add harshsinghmp/skills

2. Install ALL skills in one command

npx skills add harshsinghmp/skills -a

3. Install globally (available to every AI agent on your machine)

npx skills add harshsinghmp/skills -g

Installs into ~/.agents/skills/ so Claude Code, Codex, OpenCode, Antigravity, and other local agents can access every skill across all your projects.

4. Install specific skills directly

# Workflow & Reasoning
npx skills add harshsinghmp/skills/systematic-debugging
npx skills add harshsinghmp/skills/save-progress
npx skills add harshsinghmp/skills/writing-plans

# Engineering & Architecture
npx skills add harshsinghmp/skills/typescript-pro
npx skills add harshsinghmp/skills/sql-pro
npx skills add harshsinghmp/skills/mcp-builder

# Agency Divisions (marketing, engineering, governance)
npx skills add harshsinghmp/skills/agency-brand-guardian
npx skills add harshsinghmp/skills/agency-backend-architect

# Marketing & Social Media
npx skills add harshsinghmp/skills/hook-generator
npx skills add harshsinghmp/skills/post-writer

# Operations & Life Automation
npx skills add harshsinghmp/skills/inbox-executive-assistant
npx skills add harshsinghmp/skills/crm-hygiene

# Cognitive & Memory
npx skills add harshsinghmp/skills/muse-ground
npx skills add harshsinghmp/skills/save-progress

Compatible agents: Claude Code · Codex CLI · Antigravity · OpenCode · Oh-My-Pi · Hermes · Cursor · Windsurf · any MCP client.

📚 Skill Catalog

142+ curated AI agent skills organized into 7 categories:

Category	Count	Representative Skills	What You Get
🛡️ Quality, Security & Governance	11	`agency-code-reviewer`, `secure-code-guardian`, `security-reviewer`, `exposure-audit`	Verification gates, credential-exposure audits, OWASP-aligned security review
⚡ Engineering & Architecture	33	`typescript-pro`, `javascript-pro`, `sql-pro`, `database-optimizer`, `spec-miner`, `mcp-builder`, `agency-backend-architect`	System design, type safety, query tuning, MCP server authoring
🎨 Creative, UI/UX & Visuals	19	`agency-brand-guardian`, `show-me`, `gemini-carousel`, `graphic-designer`, `ux-patterns`	Design systems, visual rendering, accessibility, motion design
🚀 Growth, Marketing & Social	45	`hook-generator`, `post-writer`, `newsletter-voice`, `profile-optimizer`, `analytics-dashboard`	Viral hooks, platform-specific content, audience research
📦 Operations & Delivery	17	`inbox-executive-assistant`, `crm-hygiene`, `meeting-notes`, `invoice-chaser`, `weekly-review`	Admin automation, CRM hygiene, calendar and follow-up workflows
🔄 Orchestration & Paid Media	8	`agency-agents-orchestrator`, `agency-paid-social-strategist`, `agency-ppc-campaign-strategist`, `agency-tracking-measurement-specialist`	Pipeline orchestration, paid-media strategy, measurement and attribution
🧠 Cognitive, Knowledge & Memory	9	`muse-ground`, `muse-capture`, `muse-graph`, `muse-wiki`, `save-progress`	Persistent agent memory and structured second-brain workflows

Every skill ships with a SKILL.md featuring YAML frontmatter (name, description) that any modern AI agent can index and load. Disciplined developer workflows (test-driven-development, systematic-debugging, writing-plans, executing-plans, using-git-worktrees) live in the Engineering category.

🎯 Why an Aggregator?

Skills for AI agents are authored across dozens of separate repositories. A single-source hub solves three real problems:

One install command — npx skills add harshsinghmp/skills instead of remembering a dozen GitHub URLs.
Always current — a nightly sync engine pulls the latest versions from every upstream, so skills update without you lifting a finger.
Provenance & trust — every synced skill carries an immutable .upstream-meta.json stamp (source repo, branch, commit SHA, sync time), so you always know exactly which upstream commit produced a prompt.

🔄 Automated Upstream Synchronization

A GitHub Actions Sync Engine (.github/workflows/sync-skills.yml) runs daily at 04:00 UTC and performs four steps:

Discover — re-scans all 12 upstream repos and registers skills newly added upstream (zero manual manifest edits).
Sync — pulls changed skills using one shallow sparse clone per upstream, concurrent workers, retry logic, and default-branch fallback.
Validate — checks YAML frontmatter and skill structure (bun run validate).
Review PR — opens an auto-mergeable pull request with the full changelog diff.

┌─────────────────────────────────────────────────────────────┐
│                    Upstream Repositories                    │
│  obra/superpowers · harshsinghmp/muse-agents                │
│  harshsinghmp/musememory · harshsinghmp/muse-skills         │
│  ...any GitHub or git-host repo, added by URL               │
└──────────────────────────────┬──────────────────────────────┘
                               │  Daily GitHub Action (04:00 UTC)
                               │  discover → sync → validate → PR
                               ▼
┌─────────────────────────────────────────────────────────────┐
│             harshsinghmp/skills (Downstream Hub)            │
│  • Auto-discovers new upstream skills (--discover)          │
│  • Content-hash sync: zero churn when nothing changed       │
│  • Provenance stamps (.upstream-meta.json)                  │
└─────────────────────────────────────────────────────────────┘

CLI Commands

bun install

# Sync all upstreams defined in skills.manifest.json
bun run sync

# Sync + re-scan upstream repos to register newly added skills
bun run sync -- --discover

# Sync a specific skill by name filter
bun run sync -- --filter answer-first

# Preview discovery + sync without writing anything
bun run sync -- --discover --dry-run

# Validate all skills conform to the npx skills standard
bun run validate

# Run the test suite
bun test

➕ Adding Skills by Link (Aggregator Ingest)

Turn any upstream repository into part of this hub with a single command. The ingest engine resolves the link, discovers every skill inside it, registers it in skills.manifest.json, and syncs it immediately:

# Whole repo: ingest EVERY skill found in it
bun run add https://github.com/obra/superpowers

# Single skill via a tree URL
bun run add https://github.com/obra/superpowers/tree/main/skills/test-driven-development

# Shorthand works too
bun run add obra/superpowers
bun run add obra/superpowers/skills/test-driven-development

# Non-GitHub git hosts (GitLab, Bitbucket, self-hosted) via pure git protocol
bun run add https://gitlab.com/owner/repo

# Options
bun run add <url> --category engineering   # force a category
bun run add <url> --no-sync                # register without syncing
bun run add <url> --dry-run                # preview what would be registered

Input	Behavior
`https://github.com/owner/repo`	Discover & ingest all skills in the repo
`https://github.com/owner/repo/tree/<ref>/<path>`	Ingest that one skill
`https://github.com/owner/repo/blob/<ref>/<path>/SKILL.md`	Ingest that one skill
`https://raw.githubusercontent.com/...`	Ingest that one skill
`owner/repo`	Discover & ingest all skills
`owner/repo/<path>`	Ingest that one skill
Any other git host URL	Clone & scan via git protocol

Every ingested skill keeps its upstream link in the manifest — all future bun run sync -- --discover runs keep it updated automatically. Name collisions are suffixed (-2, -3), never silently overwritten.

📜 Declarative Manifest (`skills.manifest.json`)

Prefer manual control? Add an entry directly to skills.manifest.json:

{
  "name": "my-new-skill",
  "category": "engineering",
  "description": "Clear actionable description for LLM indexing",
  "upstream": {
    "repo": "upstream-org/upstream-repo",
    "branch": "main",
    "sourcePath": "skills/my-new-skill"
  }
}

Then run bun run sync to pull it with automatic provenance stamping.

Repository architecture: see docs/ARCHITECTURE.md for the full syndication spec — sparse-checkout mechanics, discovery engine, and provenance contract.

❓ FAQ

Which AI agents can use these skills? Any agent that reads SKILL.md files: Claude Code, Codex CLI, Antigravity, OpenCode, Cursor, Windsurf, Hermes, and MCP-based clients via the npx skills standard.

How do skills stay up to date? A nightly GitHub Action discovers and syncs changes from 12 upstream repositories. Upstream edits reach this hub within 24 hours, gated behind a validated review PR.

Can I add my own skills? Yes — pass any repo URL to bun run add <url>, or open a PR editing skills.manifest.json.

Are the original skill authors credited? Yes. Every skill directory carries .upstream-meta.json with its source repository and exact commit, preserving full provenance.

Is this free? MIT licensed. Install, fork, and redistribute freely.

Name		Name	Last commit message	Last commit date
.github/workflows		.github/workflows
docs		docs
scripts		scripts
skills		skills
src		src
tests		tests
.env.example		.env.example
.gitignore		.gitignore
AGENTS.md		AGENTS.md
LICENSE		LICENSE
README.md		README.md
bun.lock		bun.lock
package.json		package.json
skills.manifest.json		skills.manifest.json
tsconfig.json		tsconfig.json

Navigation Menu

Repository files navigation

🧠 AI Agent Skills Hub — 142+ Production-Grade Skills for Claude Code, Codex & MCP Agents

📑 Table of Contents

⚡ Quick Start — Install Skills with `npx skills`

1. Interactive Multi-Select (pick the skills you need)

2. Install ALL skills in one command

3. Install globally (available to every AI agent on your machine)

4. Install specific skills directly

📚 Skill Catalog

🎯 Why an Aggregator?

🔄 Automated Upstream Synchronization

CLI Commands

➕ Adding Skills by Link (Aggregator Ingest)

📜 Declarative Manifest (`skills.manifest.json`)

❓ FAQ

📄 License

About

Releases

Packages

Contributors

Languages

Latest commit

History

Folders and files

Repository files navigation

🧠 AI Agent Skills Hub — 142+ Production-Grade Skills for Claude Code, Codex & MCP Agents

📑 Table of Contents

⚡ Quick Start — Install Skills with npx skills

1. Interactive Multi-Select (pick the skills you need)

2. Install ALL skills in one command

3. Install globally (available to every AI agent on your machine)

4. Install specific skills directly

📚 Skill Catalog

🎯 Why an Aggregator?

🔄 Automated Upstream Synchronization

CLI Commands

➕ Adding Skills by Link (Aggregator Ingest)

📜 Declarative Manifest (skills.manifest.json)

❓ FAQ

📄 License

About

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages

⚡ Quick Start — Install Skills with `npx skills`

📜 Declarative Manifest (`skills.manifest.json`)