onboardingV2
Warn
Audited by Socket on Jun 3, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The core behavior matches LaunchDarkly onboarding, and most data flows point to official LaunchDarkly services and official SDK packages. Risk comes from transitive skill installation via `npx skills add`, especially the less-verifiable `launchdarkly/experimental-agent-skills`, plus automatic credential handling and silent file writes. This looks coherent with its purpose but carries medium supply-chain and delegated-execution risk.
Confidence: 87%Severity: 68%
Audit Metadata