importing-chatgpt-memory

SUSPICIOUS: the skill is broadly aligned with its stated purpose of ChatGPT-to-Letta memory onboarding and shows no external exfiltration, credential harvesting, or suspicious installer behavior. However, it handles very sensitive archives, encourages broad default retention of personal details, and combines large-scale ingestion of untrusted conversation content with autonomous file-writing and subagent mining, creating meaningful privacy and prompt-injection risk even though the overall workflow remains locally scoped.