@349/okx-onchainos-suite

Warn

Audited by Socket on Mar 18, 2026

4 alerts found:

Securityx3Anomaly
SecurityMEDIUM
modules/okx-agentic-wallet/SKILL.md

SUSPICIOUS. The skill’s purpose and capabilities mostly align for a wallet agent, and the installer path is same-org and checksum-verified. However, it is inherently high risk because it lets an AI agent authenticate a wallet, handle sensitive credentials, and execute irreversible on-chain transfers and contract calls; cross-skill routing further broadens trust. This is not confirmed malware, but it is a high-risk financial-action skill.

Confidence: 90%Severity: 81%
SecurityMEDIUM
SKILL.md

The skill’s stated purpose is coherent with OKX OnchainOS workflows, so this does not look like confirmed malware. However, it requires an unverifiable external `onchainos` CLI and sends high-value OKX credentials to that binary, which triggers a high-risk classification under the dependency and credential-forwarding rules. Overall verdict: SUSPICIOUS/HIGH RISK due to opaque external execution and credential trust, not due to purpose mismatch.

Confidence: 86%Severity: 86%
SecurityMEDIUM
modules/okx-dex-swap/SKILL.md

SUSPICIOUS: the skill is purpose-aligned and uses same-org distribution paths, but it grants an AI agent the ability to approve and execute cryptocurrency trades, including optional automated mode. That real-world transaction capability makes the overall risk high even without clear evidence of malicious intent or hidden exfiltration.

Confidence: 87%Severity: 78%
AnomalyLOW
modules/okx-dex-trenches/SKILL.md

SUSPICIOUS. The main memecoin research capability aligns with the stated purpose, and the installer source is same-org OKX GitHub with checksum verification, which lowers concern. But the skill still relies on remote script execution, expands into transitive wallet/swap workflows with real financial impact, and processes untrusted external token content while having command execution capability. No direct credential theft or clear malicious exfiltration is evident.

Confidence: 81%Severity: 56%
Audit Metadata
Analyzed At
Mar 18, 2026, 03:48 PM
Package URL
pkg:socket/skills-sh/Starchild-ai-agent%2Fcommunity-skills%2Fokx-onchainos-suite%2F@34baf2dc49d30f5932779d2110de5db846bca93f