Skip to main content
Glama
auuduu

dsh-timem-memory

by auuduu

dsh-timem-memory

A DeepSeek Harness (DSH) plugin for TiMEM long-term memory. One cordis entry sets up two things:

  1. MCP bridge — connects to the TiMEM cloud MCP via @deepseek-ai/dsh-mcp-client (https://api.timem.cloud/mcp), registering 17 tools for the model under the mcp__timem__* namespace (4 memory, 8 rules, 4 knowledge base, 1 ready);

  2. Memory skill — registers 5 built-in skills into the DSH skill system (runtime layer, rank 250), so the model automatically executes the workflow "search memory first each round → do work → write memory", without requiring the user to understand S0–S6 levels.

Credentials are resolved via ctx.credentials (TiMEM_API_KEY / TiMEM_USER_ID), no keys are hardcoded in any configuration file; credential changes trigger an event to automatically re-mount the MCP bridge without restart.

Installation

Prerequisites: DSH is installed, pnpm is on the PATH (dsh plugin depends on it).

# 1. 把插件装进目标 profile(任意名字,web / headless / 自定义皆可)
#    GitHub 渠道(当前即可用):
dsh plugin --profile web add git+https://github.com/auuduu/dsh-timem-memory.git
#    npm 官方商店渠道(发布后):
# dsh plugin --profile web add dsh-timem-memory

# 2. 在 <DSH_HOME>/profiles/web/cordis.patch.yml 中插入一条目(保存即热生效)
- insert:
    - id: timem-memory
      name: 'dsh-timem-memory'

Simple explanation for beginners (what the package is, where installation goes, how DSH "activates" it): docs/principles.md

Related MCP server: r3 (Recall)

Configure Credentials

Credentials are stored in the credential store (~/.dsh/.credentials.yaml, permission 0600, manageable via the DSH web credential page), or in environment variables / ~/.dsh/.env (loaded at startup). Using .credentials.yaml as an example:

DEEPSEEK_API_KEY: sk-…
TiMEM_API_KEY: sk-…          # https://console.timem.cloud/settings/api-key
TiMEM_USER_ID: usr_…         # TiMEM 控制台个人资料页

After writing, the plugin will automatically re-mount the MCP bridge (no need to restart DSH). If TiMEM credentials are missing, the plugin only skips the bridge and logs an error; skills are still registered.

Configuration Options

Field

Default

Description

serverName

timem

MCP namespace, tool name seen by model: mcp__<serverName>__<tool>

url

https://api.timem.cloud/mcp

TiMEM cloud MCP endpoint

apiKeyRef

TiMEM_API_KEY

API Key reference name in ctx.credentials

userIdRef

TiMEM_USER_ID

User ID reference name in ctx.credentials

enableSkills

true

Whether to register built-in skills

skillNames

5 built-in skills

List of skills to register

toolCallTimeoutMs

60000

Pass-through to MCP client

failOnStartupError

false

Pass-through to MCP client

reconnect.*

Client defaults

Pass-through to MCP client

Verification

After restart or hot-reload, ask the model in a conversation "Is TiMEM connected?" — it will call mcp__timem__ready and confirm that 17 tools are visible; or check the status of the timem-memory entry in the Settings → Plugins page.

How It Works

  • lib/index.js: the cordis plugin itself (inject: ['credentials', 'skills']). On activation: ① resolve credentials → mount MCP bridge with ctx.plugin(mcpClient, …); ② read skills/*/SKILL.md from the package → ctx.skills.register() (auto-destroyed with context).

  • skills/: 5 skill packages (DSH-adapted versions of official TiMEM skills, tool names rewritten to mcp__timem__*, with built-in availability checks and S3/S0/S6 levels).

  • Credential hot reload: listens for credentials/updated event, when apiKeyRef / userIdRef matches, releases the old bridge and re-mounts with new credentials.

License

MIT

A
license - permissive license
-
quality - not tested
C
maintenance

Maintenance

Maintainers
Response time
Release cycle
Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

  • A
    license
    -
    quality
    D
    maintenance
    A fully self-hosted MCP server that integrates the Mem0 framework to provide persistent memory capabilities for AI assistants using local models and vector storage. It enables users to store, search, and manage contextual information across conversations through a Docker-based deployment.
    3
    MIT
  • A
    license
    B
    quality
    C
    maintenance
    An intelligent memory MCP server that provides AI applications with semantic search, entity extraction, and knowledge graph capabilities using local Redis caching and optional cloud sync. It enables LLMs to store and retrieve long-term context across sessions with high-performance multi-tier storage.
    12
    11
    2
    MIT
  • A
    license
    A
    quality
    A
    maintenance
    MCP server providing persistent memory and context for AI tools, including semantic memory, knowledge graph, and session history to avoid starting from scratch in every conversation.
    34
    14
    MIT
  • A
    license
    -
    quality
    C
    maintenance
    MCP server that provides AI agents with persistent memory, cross-agent sharing, and context management, enabling them to remember conversations, track complex tasks, and evolve skills across tools.
    2
    MIT

View all related MCP servers

Related MCP Connectors

  • Cloud-hosted MCP server for durable AI memory

  • Your memory, everywhere AI goes. Build knowledge once, access it via MCP anywhere.

  • Persistent memory and cross-session learning for AI coding assistants (hosted remote MCP).

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/auuduu/dsh-timem-memory'

If you have feedback or need assistance with the MCP directory API, please join our Discord server