The v0.0.2 release of Knowledge-and-Memory-Management introduces a deploy-agnostic update for knowledge-driven agents. It replaces hardcoded personal paths with the portable $AGENT_HOME variable, removing environment-specific configurations. The update focuses on collecting knowledge from web pages, videos, and articles. Memory management is built around flexible and portable storage solutions. The web collector uses adaptive scraping with headless browser support and respects robots.txt. Video knowledge is extracted via transcript analysis, preserving metadata. Article extraction focuses on long-form content, normalizing it into a structured format. All collectors return a uniform data object for indexing and storage by the memory system. Memory is stored in two layers: raw content and processed vector indexes, both under $AGENT_HOME. Deduplication uses content hashes, and nightly tasks prune outdated entries, simplifying backup and migration.