dsh-engram-session
Run the following command in DeepSeek Harness:
dsh plugin install xiuyuan18/dsh-engram-session
Paste the following prompt into your AI chat to install this plugin:
Install the plugin by running dsh plugin install xiuyuan18/dsh-engram-session in DeepSeek Harness; the source code is available at https://github.com/xiuyuan18/dsh-engram-session
About this plugin
Running multiple concurrent sessions in DeepSeek Harness exposes two real pain points with Engram memory: every session shares the same global tool layer, so memory entries bleed across unrelated contexts, and hosts such as the DSH web GUI, VS Code, WSL, or CI pipelines cannot reliably pass a working directory, which breaks Engram's project auto-detection. dsh-engram-session is a standalone bundle that resolves both issues, giving each session its own isolated and deterministic memory surface.
Per agent session the plugin spawns a dedicated Engram MCP stdio child rooted at that session's workspace, then registers the mem_ tools in the agent's own scope via agent.ctx.tools.register(). Because scoped registration shadows the global tool layer, concurrent sessions never collide even though they share the same mcp__engram__mem_ names. Binary resolution is fully automatic, walking PATH and common install locations in order; if no executable is found the plugin degrades gracefully and simply skips the protocol section, so an agent is never shown tools that cannot exist. The plugin also patches the scope parameter descriptions shipped by Engram (adding the missing global value) so the agent's tool contract stays accurate without waiting for an upstream release.
This plugin is best for developers who run parallel DSH sessions, switch between local GUIs and VS Code extensions, work inside WSL containers or CI pipelines, or simply want memory to follow the project repository rather than living in a single global store. It never modifies the deployment persona or the system prompt, and uninstalling the bundle restores the exact previous state, making it a minimal-diff, fully reversible memory deployment ideal for teams that value clean rollbacks.
Use Cases
- Preventing memory bleed across concurrent sessions
- Fixing Engram project auto-detection in GUI or WSL hosts
- Reusing the same memory bundle across multiple profiles
Best For
- Developers running multiple DSH sessions in parallel
- Engineers using DSH via VS Code, WSL, or CI pipelines
- Teams that prefer repository-scoped over global memory
Related Plugins
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.
An opinionated, zero-infrastructure file-based memory protocol: human-readable markdown files plus a bounded always-loaded index, curation discipline, and no database or embeddings.