What problem does it solve? Moving an existing project from the self-hosted mem0 OSS SDK (the local Memory class with vector store, embedder, and LLM config) to the hosted mem0 Platform (MemoryClient) involves subtle API differences, config removal, and behavioral changes that are easy to get wrong. This Skill audits every mem0 touchpoint in a repo, writes a reviewable migration plan, and executes it only after approval. ## Core Features & Use Cases - Full Footprint Discovery: Finds every mem0 import, client initialization, config block, call site, dependency, env variable, and local infrastructure service across Python and TypeScript codebases. - Verified API Mapping: Confirms real method signatures against the installed SDK version, then maps each OSS call to its hosted equivalent (filters, top_k/topK, pagination) using a detailed reference mapping. - Plan-Then-Execute Workflow: Writes a concrete MEM0_MIGRATION_PLAN.md with inventory, change set, concerns, and verification steps, then stops for developer approval before editing any code. - Use Case: A team running a RAG service with Memory.from_config backed by a local Qdrant container wants to switch to the managed platform. The skill inventories all call sites, flags that existing local memories will not carry over, and migrates the code to MemoryClient with a smoke test against the real API. ## Quick Start Ask your assistant to migrate this project's mem0 setup from the self-hosted OSS SDK to the hosted mem0 Platform using my MEM0_API_KEY.