dsh-memory：DSH 两层渐进式长期记忆
dsh-memory 是 jipika 为 DeepSeek Harness 维护的记忆类插件：全局 topics 分片与按工作区切换的项目层均以纯 markdown 存储，系统提示词仅注入实时片级地图，条目通过 memory_search 与 read 按需加载。插件不调用模型，支持设置面板编辑、三档注入方式与写入守卫。安装命令为 dsh plugin --profile web add @jipika/dsh-memory。
Read MoreDeepSeek Harness 插件：dsh-memory（跨会话记忆）
DeepSeek Harness 的 `dsh-memory` 插件通过插件化架构，填补了现有智能体开发中长对话记忆断层的空白。该插件核心功能包括：会话结束时自动总结对话内容、新会话开始时自动注入过往笔记，以及将数据本地存储在 `$DSH_HOME/memory/` 目录下。 它支持通过环境变量进行无UI配置，用户可灵活控制蒸馏频率、召回范围及文件大小等参数。该插件数据安全，不依赖会话日志，仅提供自动召回与自动蒸馏两条路径。
Read Moredsh-memory: DSH Advanced Auto-Memory Plugin
The dsh-memory plugin introduced by DeepSeek Harness (DSH) aims to address context window limitations in agent development. Built on local SQLite storage, it provides core features such as tiered memory, vector semantic retrieval, knowledge graph visualization, and worldline version management, supporting automatic memory injection, consolidation, and governance. In addition, the plugin offers advanced capabilities such as LLM distillation and global deduplication, and exposes the memory_* series of tool commands. Once installed, it can be enabled in a DSH environment and requires Node.js >= 22.19.
Read Moredsh-memory: A Cross-Session Long-Term Memory Plugin for DSH
`dsh-memory` is a local memory plugin for DeepSeek Harness (DSH) designed to address models' lack of cross-session memory. Built on local file-system storage and operating without LLM calls, it uses pure string matching and hash-based deduplication to enable zero-cost long-term memory persistence. Core features include multi-dimensional retrieval, conflict detection with a human adjudication mechanism, and audit logs. It injects memory into context through an L1, L2, and L3 three-tier structure, allowing the model to read and write memory on demand via tool calls. The plugin also provides a settings panel to manage memory, view usage, and resolve conflicts, making it well suited for workflows that require maintaining project context and preferences.
Read Moredsh-memory: A Long-Term Workspace Memory Plugin for DeepSeek Harness
The dsh-memory plugin is designed to address memory management challenges in DeepSeek Harness by providing workspace-based long-term memory capabilities. Its core features include indexed briefings to reduce token consumption, relevant recall at the end of each turn, version-chain management, and automatic distillation. Users can perform operations such as saving and searching through the `memory` tool, and manage the memory store in the settings panel. Installation requires using a PowerShell script and ensuring configuration consistency. Data is stored locally by default, and a repository mirror mode can be enabled. The plugin is maintained by BangBang-03 and released under the MIT license.
Read Moredsh-memory: Persistent Memory Plugin for DeepSeek Harness
The `dsh-memory` plugin for DeepSeek Harness (DSH) is designed to address the loss of context after an agent session restart. It implements persistent storage using pure Markdown files and supports user-level and project-level memory isolation, ensuring agent consistency across different sessions. Its core features include cross-session memory recall, a visual memory manager, and project memory isolation. When using it, note the ordering of the `--profile` parameter and restart the service to register the tool. The plugin has zero dependencies and ensures safety through atomic writes and boundary design, making it well suited for agent development scenarios requiring long-term context management.
Read Moredsh-memory: Localized Long-Term Memory for DeepSeek Harness
`dsh-memory` is a plugin for DeepSeek Harness that addresses the problem of agents forgetting user preferences and project constraints after a regular session ends. It integrates memory organization approaches from Claude Code and Codex, providing core capabilities such as local Markdown storage, global and workspace isolation, progressive disclosure, and secure writes. Users can enable it with an installation command, configure the memory organization model in the Web interface, and achieve long-term memory management across projects. The plugin is open source under the MIT license, ensuring data privacy and a local-first approach.
Read MoreDeepSeek Harness Memory Plugin: Persistent, Searchable Project Context
`dsh-memory` is a persistent memory plugin for DeepSeek Harness (DSH), designed to address the challenge of agents maintaining context and rule consistency across long conversations. The plugin stores decisions, rules, and session summaries in a local DuckDB database and provides a web interface for management, retrieval, and import/export. Its core features include memory management with semantic and keyword search, rule enforcement through system prompt injection, session context maintenance for top-level agents, and support for custom data lifecycle policies. The plugin requires Node.js version 22.19 or higher and allows rules to be imported from Markdown files. This solution significantly improves the maintainability and consistency of agent projects.
Read Moredsh-memory: A Cross-Session Semantic Memory Plugin for DeepSeek Harness
`dsh-memory` is a cross-session semantic memory plugin for DeepSeek Harness, designed to solve the problem of agents losing context after a session ends. It uses SQLite and Zhipu Embedding-3 to perform local vector storage and retrieval, and can be enabled in a configuration file without requiring new infrastructure. The plugin provides three core tools: `memory_add` for persisting facts, `memory_search` for semantic retrieval with keyword fallback, and `memory_update` for correcting memories while preserving their history. Memory can be managed via the command `/mem <question>` or by using configured trigger phrases (such as “remember”). The plugin supports graceful degradation, automatically falling back to keyword search when services are unavailable or network conditions are poor. After installation, a restart is required.
Read Moredsh-memory: A Persistent Atomic Fact Memory Plugin for DeepSeek Harness
The `dsh-memory` plugin for DeepSeek Harness (DSH) is designed to address the loss of context in multi-turn agent conversations. Built on Cordis, it uses local JSON documents for persistent storage and does not rely on external databases. Its core is the atomic fact model (Subject-Predicate-Object), which uniformly manages semantic, episodic, and procedural memory. The plugin provides tools such as `recall`, `remember`, and `forget`, supports capturing user preferences through trigger phrases, and includes PII privacy protection mechanisms. It also supports `user.md` file synchronization and dynamic context injection, facilitating manual intervention and system reconstruction logging. Although it currently only supports BM25 semantic retrieval and LLM extraction is disabled by default,
Read Moredsh-memory: Adding Persistent AI Memory to DeepSeek Harness
`dsh-memory` is a static profile plugin for DeepSeek Harness that addresses the stateless session problem and gives AI persistent memory capabilities. Its core features include memory tab management, automatic naming, multi-AI privacy isolation, experience archiving, and AI name display in the sidebar. To install the plugin, place the repository in the designated directory and modify the configuration, then restart to activate the changes. The plugin supports writing different types of memories (such as global, shared, and experience) via the `memory_record` tool and injects guidance into the SystemPrompt to instruct AI on memory management, making it suitable for developers who need long-term context and privacy isolation.
Read Moredsh-memory: A Long-Term Memory Plugin for DeepSeek Harness
dsh-memory is a long-term memory plugin for DeepSeek Harness (dsh), designed to address the context window limitations of agents and the fragmentation issue in conversational memory. It uses persistent JSON storage, automatically guides the model to call `memory_recall` before responding to retrieve relevant background information, and calls `memory_save` after responding to store new facts. The plugin provides five core tools (create, read, update, delete, and search), supports a 1-5 star importance rating, and includes a visual web-based management interface, intelligent search, and import/export capabilities. It can inject key memories into the context, effectively improving the memory coherence and interaction quality of agents.
Read Moredsh-memory (Cairn): Persistent Memory for DeepSeek Harness
`dsh-memory` is a memory plugin for DeepSeek Harness (DSH), designed to address context loss after agent sessions end. It provides persistent cross-session storage with three-tier memory scopes: global, project, and user. The plugin leverages the BM25 algorithm for highly relevant retrieval and supports summary-first injection as well as step-level automatic recall. It also features automatic identity updates, security scanning, a deduplication pipeline, and two-tier lifecycle management, along with a management interface and manual review mechanism, making it suitable for complex task development that requires long-term context maintenance.
Read Moredsh-memory: A Persistent Three-Tier Memory Plugin for DeepSeek Harness
dsh-memory is a host-side plugin for DeepSeek Harness (DSH), designed to address the challenges of memory sharing and persistence for agents in multi-turn conversations and cross-workspace scenarios. By introducing a three-tier memory structure (long-term, short-term, and session summaries), it enables automatic memory accumulation and maintenance. It uses system prompts to guide all preconfigured agents to proactively invoke memory tools, and provides capabilities such as `memory_read`, `write`, and `search`. After installation, the system automatically monitors conversation events, stores raw logs and LLM-generated summaries in separate directories, and ensures that memories remain available across sessions.
Read Moredsh-memory: Persistent Memory Storage for DeepSeek Harness
`dsh-memory` is a community plugin for DeepSeek Harness designed to address the lack of persistent cross-session memory for agents. The plugin enables agents to store, retrieve, delete, and search facts using stable key-value pairs, with tag support. It provides four core tools: `memory_set`, `memory_get`, `memory_delete`, and `memory_search`. Installation is simple and requires no build step, relying on a storage backend provided by the environment, such as SQLite or JSON. This tool is well suited for scenarios where specific state or a knowledge base must be maintained across multi-turn conversations.
Read Moredsh-memory: Shared Persistent Memory Plugin for DeepSeek Harness
`dsh-memory` is a plugin designed for DeepSeek Harness (DSH) to address the issue of agents being unable to retain context across sessions. By creating a shared persistent memory store at `$DSH_HOME/storages/dsh-memory.json`, it enables read/write synchronization across all session types, including Web and headless sessions. The plugin provides four core tools: `memory_add`, `memory_search`, `memory_list`, and `memory_remove`, with support for tag and category management. Agents can use these tools to proactively store or retrieve information, and the plugin can also be operated directly through the sidebar of the Web interface. In addition, it uses an atomic write mechanism.
Read Moredsh-memory: A Codex-Style Persistent Memory Plugin for DeepSeek Harness
The DeepSeek Harness plugin `dsh-memory` provides persistent memory for agents, addressing the loss of context in multi-turn conversations. Its core features include persistent storage, automatic summary injection, a memory toolset, and configuration via a web interface. Installation requires specifying a version hash and restarting the environment. The plugin operates in the `$DSH_HOME/memories/` directory and achieves long-term memory capabilities similar to Codex by automatically distilling conversation turns and periodically merging them into a global summary. It supports initializing seed memory from `AGENTS.md` and provides tools for reading and writing memory.
Read Moredsh-memory: Cross-Session Memory and Vector Retrieval Plugin for DeepSeek Harness
`dsh-memory` is a cross-session memory plugin designed for DeepSeek Harness, aimed at resolving the loss of context after agent restarts. Maintained by developer zhouzhencheng07, it features core capabilities including automatic background memory, maintaining daily workspace notes, memory retrieval, and optional vector search. The plugin runs with the permissions of the current process; users are advised to review its source code and license before installation. It provides an effective solution for handling long-term memory in the DSH ecosystem.
Read Moredsh-memory: Local SQLite Cross-Session Memory for DeepSeek Harness
`dsh-memory` is a lightweight cross-session memory plugin for DeepSeek Harness (DSH), released under the MIT License. It implements memory persistence using local SQLite files and FTS5 indexes, eliminating the need for dependencies on embedded services, API keys, or external processes. It is compatible with Node.js 22.19+ or 24.0+. Its core functionality includes three model tools: `memory_write` (for storing labeled persistent facts), `memory_search` (for keyword-based lexical retrieval), and `memory_forget` (for deleting outdated memories). Additionally, the plugin provides the `memory:recall` prompt segment, which renders memories in a "pinned priority, recently updated" order within the character budget.
Read Moredsh-memory: Long-term Memory Plugin for DeepSeek Harness
dsh-memory is an MIT-licensed long-term memory plugin for DeepSeek Harness (DSH), designed to address the lack of a stable medium for user preferences, verification steps, and failure lessons in multi-session workflows. Distributed via GitHub, it utilizes structured extraction to convert reusable information from completed sessions into summaries and memory files, and automatically injects them into subsequent sessions. Core features include: debounced extraction after session completion (default 3 minutes), global cooldown integration (default 6 hours) into MEMORY.md and memory_summary.md, persistent summary injection, and the provision of four query and add tools: memory_list, memory_read, memory_search, and memory_add. Among them, memory_add is only used when...
Read Moredsh-memory: Memory with citations based on DSH session logs
`dsh-memory` is a plugin based on DeepSeek Harness (DSH) designed to address memory distortion and tracing difficulties in multi-turn agent conversations. Leveraging DSH's lossless conversation logs, it converts persistent facts into human-auditable Markdown memory files through offline distillation after sessions conclude, storing them in the `~/.dsh/memory` directory. The core advantage lies in every memory being accompanied by a citation, allowing for precise tracing back to original conversation log events; it supports reading content via `memory_read` and viewing original excerpts via `memory_expand`. The plugin provides a compact Recall Index to optimize token usage and supports `user`, `project`, `f...`
Read MoreDSH Cross-session Memory Plugin dsh-memory: YAML Storage, Semantic Retrieval, and Two-level Injection
`dsh-memory` is an MIT-licensed cross-session memory plugin developed specifically for DeepSeek Harness (dsh), designed to address context loss issues in new sessions, such as the need to repeatedly explain project preferences and historical learnings. The plugin stores memories via YAML files (default `memory.yaml`) and provides six core tools—`remember`, `recall`, `view`, `list`, `forget`, and `pin`—to implement memory saving, mixed retrieval (keywords + local Ollama embedding), viewing, and lifecycle management. Its main features include adopting a "title-first" injection strategy to reduce token usage; and supporting `session` (fixed injection) and `auto` (on-demand injection).
Read More