agent-reach
Warn
Audited by Socket on Oct 11, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The skill’s broad internet-access purpose matches many listed capabilities, but its actual footprint is too trust-heavy: mutable GitHub ZIP installation, execution of opaque bundled binaries, browser cookie extraction, and account-action automation. Data flows are not clearly limited to official first-party APIs, and the credential scope is disproportionate for a generic channel-enablement skill.
Confidence: 91%Severity: 84%
Audit Metadata