Show HN: File-base agent memory framework, like Claude's skills (opens in new tab)

AI memory systems often become a black box. When an LLM produces a wrong answer, it’s unclear whether the issue comes from storage, retrieval, or the memory itself.

Most systems rely on RAG and vector storage, which makes memory opaque and hard to inspect, especially for temporal or multi-step reasoning.

An alternative is to make memory readable and structured: store it as files, preserve raw inputs, and allow the LLM to read memory directly instead of relying only on vector search.