openclaw

Warn

Audited by Socket on Mar 9, 2026

1 alert found:

Security
SecurityMEDIUM
SKILL.md

The OpenClaw skill framework documentation is coherent with its stated purpose of designing and distributing local AI assistants and skills, including production hardening. However, the presence of install.sh via curl|bash and install.ps1 patterns constitutes a significant supply-chain risk and should be treated as suspicious rather than benign. The documented requirement for environment secrets (API_KEY) is legitimate for authentication but must be tightly sandboxed. Overall, the skill set is potentially benign if used with strict provenance verification, signed artifacts, and proper sandboxing; otherwise, it poses notable security risks due to download-execute patterns and credential handling.

Confidence: 65%Severity: 75%
Audit Metadata
Analyzed At
Mar 9, 2026, 06:58 AM
Package URL
pkg:socket/skills-sh/realjaymes%2Fmarketingagentskills%2Fopenclaw%2F@80ff2f6c53b22c099d218b9bb2c63e9b9b539889