autoresearch
Warn
Audited by Socket on Mar 18, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
BENIGN in purpose alignment but HIGH-IMPACT operationally: it is a coherent local ML automation skill, not a credential harvester or exfiltration tool. Main risk comes from indefinite autonomous code modification/execution and destructive git actions; supply-chain risk is limited to official but unpinned `uv` usage.
Confidence: 89%Severity: 68%
Audit Metadata