agent-sessions
Warn
Audited by Socket on Oct 15, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The skill’s purpose and data access are broadly aligned: it reads local agent histories and can resume sessions, with no obvious credential exfiltration or third-party API proxying. However, the core functionality depends on a local absolute-path executable whose provenance is not established in the skill, and it also relies on unrelated third-party local tools. That unverifiable binary trust issue is disproportionate enough to make the skill high risk despite otherwise coherent purpose and local-only data flow.
Confidence: 89%Severity: 72%
Audit Metadata