lmzhen/dsh-evolution--packages-tool-memory1

@deepseek-ai/dsh-tool-memory

Model-facing memory tool and prompt context

AI 分析

核心用途是为模型提供记忆工具和运行时快照上下文。适合需要让模型具备持久化记忆、在请求间保持上下文连贯性的任务,快照大小随存储条目增加而变化。

パッケージ
@deepseek-ai/dsh-tool-memory
バージョン
0.3.70
ライセンス
MIT
最終更新
2026/09/12

インストール

検証済み bundle がないか、互換性チェックに失敗しています。先にリポジトリの説明を読んでください。 README 全文を読む ↗

ドキュメント

README 全文を読む ↗

@deepseek-ai/dsh-tool-memory

Model-facing memory tool and prompt context

Model Experience

Memory tool and runtime snapshot

What the model sees

The model sees the memory tool schema, one fixed guidance section, and a runtime snapshot containing the current memory entries.

Token effect

Fixed guidance has a constant token cost. The runtime snapshot scales with stored memory entries and is absent when memory is empty.

KV Cache effect

Guidance text is prefix-stable. The runtime snapshot is replaced after successful memory writes and is otherwise unchanged between requests.

Known Limitations and Deferred Work

  • No known durable consumer gaps at this time. Runtime contracts are covered by package and boundary tests.