Review a deployed OpenClaw agent in a strictly read-only way to audit runtime behavior, detect drift across key documentation files (AGENTS.md, SOUL.md, USER.md, MEMORY.md), assess memory hygiene and interaction quality, and evaluate tooling reliability and workflow design. This process generates structured recommendations for safe long-term improvement without automatically changing code, config, or production behavior.
Design e Frontend#deployby wangzhxg
