clawsmith.com/claw/snyk-toxicskills-1467-malicious-clawhub-prompt-injection
⚠ IssueCompetitiveLive
Snyk ToxicSkills: 36% of ClawHub Skills Have Prompt Injection, 1,467 Malicious Payloads
Snyk scanned 3,984 ClawHub skills and found 1,467 malicious payloads. 36% of all skills contain detectable prompt injection. 91% of confirmed malicious samples combine prompt injection with traditional malware.
Virality Score
5,000
across 3 platforms
Score Breakdown
HN
2,500
Reddit
2,000
X
500
Social Proof 2 sources
Gap Assessment
CompetitiveMarket has established players
Snyk ToxicSkills Scanner, VirusTotal integration, and SecureClaw all address this space. Competitive.
Details
Signalissue
Ecosystem—
Sources2
Platforms3
Updated2d ago
Trend→ stable
Related signals
92.8kAgents of Chaos: 38 Researchers Deploy 6 OpenClaw Agents — 11 Critical Failure Patterns in 14 Days35.6kCapability Evolver — #1 ClawHub Skill (35K Downloads) Caught Exfiltrating Data to ByteDance Feishu18.3kChina Bans OpenClaw from Government Computers and Banks — CERT Warns 'Extremely Weak Security'