ralph
Warn
Audited by Socket on Mar 23, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the skill's high-level capabilities fit its stated development-loop purpose, but the Ralph command surface is not verifiably the same as Anthropic's official plugin and the snippet omits install/provenance details. Main risk is autonomous modification of arbitrary codebases and analysis of untrusted project content, not confirmed malware or exfiltration.
Confidence: 79%Severity: 57%
Audit Metadata