agenty

Warn

Audited by Socket on Mar 4, 2026

1 alert found:

Security
SecurityMEDIUM
SKILL.md

Overall, the skill fragment shows a coherent, purpose-aligned integration workflow that leverages Membrane as a trusted middleware layer for authenticating and orchestrating Agenty actions. No malicious behavior or credential harvesting patterns are evident. The data flow remains within Membrane-mediated proxies, reducing risk of leakage. Consider ensuring proper access controls on connection IDs and action IDs and auditing Membrane-proxied requests for sensitive payloads.

Confidence: 75%Severity: 75%
Audit Metadata
Analyzed At
Mar 4, 2026, 08:56 AM
Package URL
pkg:socket/skills-sh/membranedev%2Fapplication-skills%2Fagenty%2F@e051b89536a0ae59f04c1c812317d35e1fc356e4