ClawHub Layer API

The complete REST API for ClawHub — 36,000+ skills with full data the official API doesn't expose.

         

   

---

Why?

The official ClawHub API is limited — no full-text search, no security scan results, no file contents, no moderation data. To get the full picture you'd need to query their Convex database directly.

ClawHub Layer API pulls the complete catalog from ClawHub's Convex DB, caches it in MongoDB, and serves it as a clean, fast REST API.

---

Try It Now

Live API — clawhub.atomicbot.ai

API Docs - swagger

OpenAPI Spec — clawhub.atomicbot.ai/docs-json — feed it to your coding agent to unleash full API access.

---

What You Get

• 📦 Complete catalog — all 36,000+ skills with stats, versions, tags, and ownership
• 🔍 Full-text search — weighted search across slug, name, and description
• 🛡️ Security data — VirusTotal scan results and LLM-based analysis
• 🚩 Moderation info — suspicious/malicious flags, removal status, reason codes
• 📄 File contents — read any file from any skill (SKILL.md, configs, source code)
• 💬 Comments — live comment threads with user profiles
• 🔀 Fork tracking — fork/duplicate relationships between skills
• 🔄 Auto-sync — configurable cron keeps your data fresh
• ⚡ On-demand enrichment — detail data fetched and cached on first request
• 📖 Swagger UI — interactive API docs at /docs

---

Quick Start

git clone https://github.com/AtomicBot-ai/clawhub-layer-api.git
cd clawhub-layer-api
docker compose -f docker-compose.prod.yml up -d

API at http://localhost:3000 · Swagger at http://localhost:3000/docs

Or use the public instance: clawhub.atomicbot.ai/docs

Trigger initial sync:

docker compose -f docker-compose.prod.yml exec app node dist/cli sync

---

API

List Skills

GET /api/skills?page=1&limit=25&sort=downloads&dir=desc&nonSuspiciousOnly=true

Param	Values	Default
sort	downloads · stars · installs · updated · newest · name	downloads
dir	asc · desc	desc
limit	1–200	25
nonSuspiciousOnly	true · false	false

Search

GET /api/skills/search?q=summarize&limit=25

Weighted full-text search: slug (×10), name (×5), summary (×1).

Skill Detail

GET /api/skills/:slug

Returns the complete skill record: metadata, latest version, owner, file list, security analysis (VT + LLM), moderation status, fork info, and SKILL.md content. Cached in MongoDB, re-fetched if stale.

File Content

GET /api/skills/:slug/files?path=SKILL.md

Returns raw file content for any file in the skill package.

Comments

GET /api/skills/:slug/comments?limit=50

Live comments from ClawHub with user profiles.

Health

GET /health

---

How Sync Works

ClawHub Convex DB ──▶ Layer API (MongoDB) ──▶ Your App
       ▲                      │
       └── cron / on-demand ──┘

1. Bulk sync (cron) — pages through the entire ClawHub catalog via listPublicPageV4, upserting into MongoDB. ~36k skills in ~2-3 minutes.
2. Detail enrichment (on-demand) — when GET /api/skills/:slug is called, full detail is fetched from getBySlug and cached with a configurable TTL.
3. File caching — file contents are fetched from ClawHub's site API and cached in MongoDB with the same TTL.

---

Configuration

Variable	Default	Description
MONGODB_URI	mongodb://localhost:27017/clawhub-layer	MongoDB connection
CONVEX_CLOUD_URL	https://wry-manatee-359.convex.cloud	ClawHub Convex query endpoint
CONVEX_SITE_URL	https://wry-manatee-359.convex.site	ClawHub site HTTP API
SYNC_CRON	0 */3 * * *	Sync schedule (default: every 3 hours)
CACHE_TTL_HOURS	3	TTL for detail/file cache
PORT	3000	Server port

---

CLI

# Full sync (inside container)
node dist/cli sync

# From outside
docker compose -f docker-compose.prod.yml exec app node dist/cli sync

---

Development

docker compose up -d    # Starts MongoDB + app with hot-reload

---

License

ISC

---

_{© 2026 Atomic Bot · Built with ❤️ · atomicbot.ai}