ClawHub Blocklist

Community-maintained blocklist of known malicious actors and skills targeting the ClawHub/OpenClaw ecosystem.

Purpose

In late January 2026, security researchers discovered 400+ malicious skills on ClawHub, the skill marketplace for OpenClaw (formerly Clawdbot/Moltbot). These skills delivered malware including:

• Atomic Stealer (AMOS) — macOS infostealer targeting crypto wallets, browser credentials, SSH keys
• Reverse shells — Hidden backdoors in otherwise functional code
• Credential exfiltration — Stealing API keys and tokens via webhooks

This repository provides machine-readable blocklists that security tools can consume to protect users.

Files

File	Description
actors.json	Known malicious actors (ClawHub usernames, GitHub accounts)
skills.json	Known malicious skills (by slug)
iocs.json	Indicators of Compromise (IPs, domains, URLs) — coming soon

Usage

Raw URLs (for tools)

https://raw.githubusercontent.com/philskents/clawhub-blocklist/main/actors.json
https://raw.githubusercontent.com/philskents/clawhub-blocklist/main/skills.json

Example: Fetch and parse

# Get list of blocked skill slugs
curl -s https://raw.githubusercontent.com/philskents/clawhub-blocklist/main/skills.json | \
  jq -r '.skills[].slug'

# Get list of blocked actors
curl -s https://raw.githubusercontent.com/philskents/clawhub-blocklist/main/actors.json | \
  jq -r '.actors[].id'

Example: Check if a skill is blocked

SKILL="polymarket-traiding-bot"
curl -s https://raw.githubusercontent.com/philskents/clawhub-blocklist/main/skills.json | \
  jq -e --arg s "$SKILL" '.skills[] | select(.slug == $s)' && \
  echo "⛔ BLOCKED" || echo "✓ Not on blocklist"

Data Format

actors.json

{
  "id": "username",
  "platform": "clawhub|github",
  "added": "YYYY-MM-DD",
  "source": "URL to research/disclosure",
  "notes": "Brief description"
}

skills.json

{
  "slug": "skill-name",
  "author": "username or null",
  "added": "YYYY-MM-DD",
  "source": "URL to research/disclosure",
  "malware": "Malware family if known",
  "notes": "Brief description"
}

Contributing

Report a malicious skill

1. Verify the skill is actually malicious (not just suspicious)
2. Document evidence: screenshots, code snippets, behavior analysis
3. Open an issue or submit a PR with:
   • Skill slug and author (if known)
   • Link to the skill on ClawHub
   • Description of malicious behavior
   • Source/reference if available

Submit a PR

1. Fork the repo
2. Add entry to appropriate JSON file
3. Include source link to public disclosure/research
4. Submit PR with description of the threat

What NOT to submit

• Skills that are merely low-quality or buggy
• Unverified suspicions without evidence
• Skills you haven't personally analyzed
• Duplicate entries

Sources

This blocklist is compiled from:

• Koi Security — ClawHavoc Campaign
• Snyk — ToxicSkills Study
• Snyk — clawdhub Malicious Campaign
• The Hacker News — 341 Malicious ClawHub Skills
• OpenSourceMalware — Clawdbot Skills Analysis
• Community reports and independent analysis

Disclaimer

This blocklist is provided as-is for informational purposes. Inclusion on this list indicates that the skill or actor has been associated with malicious activity based on publicly available research. Always perform your own analysis before making security decisions.

License

CC0 1.0 Universal — Public Domain Dedication

This data is released into the public domain. Use it freely.