dsh-Mmem
A standalone DeepSeek Harness long-term memory plugin with governed Memory Spaces, approval workflows, and policy-gated recall.
Install
$ dsh plugin --profile web add @mistymoon/[email protected]Plugin Overview & Capabilities
AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.
Key Capabilities
- Installs as the public package @mistymoon/dsh-mmem through the DSH plugin command.
- Stores governed confirmed memories with owner, source, visibility, memory kind, and append-only revision lineage.
- Supports isolated, selective, and federated Memory Space recall patterns.
- Generates per-turn summary candidates with TTL and scheduled automatic or manual approval workflows.
- Uses local BM25 recall with policy checks for owner, authority, scope, expiry, visibility, and disclosure.
- Provides a session-bound Memory Settings interface and memory browser with directory, search, and relationship graph views.
- Supports explicit migration planning, application, rehearsal rollback, and rollback for legacy MistyMoon SQLite confirmed rows.
Useful For
- Maintaining governed long-term memory across DSH sessions.
- Separating memory for projects, identities, or narratives with distinct Memory Spaces.
- Sharing controlled read-only recall between selected DSH Workspaces.
- Reviewing candidate memories manually or through local scheduled approval.
- Migrating legacy MistyMoon confirmed-memory data into a selected Memory Space archive.
Who It Fits
- DeepSeek Harness users needing persistent, governed memory.
- DSH Workspace owners managing memory access and approval policies.
- Developers integrating the plugin into a Cordis/DSH composition.
Documented Limitations
- This is an alpha release and does not declare compatibility with DSH rc.9, later release candidates, or stable releases.
- The sidebar and graph UI were fully validated only against DSH 0.1.0-rc.8; the documented peer range is >=0.1.0-rc.7 <0.1.0.
- Default retrieval is local BM25; PageIndex, graph adapters, remote engines, embeddings, and reranking are not part of the default baseline.
- Automatic summaries remain untrusted candidates and are not fine-grained fact extraction.
- After the first relationship-confirmed archive event, older plugin versions cannot read the new archive event type, so direct downgrade reading is unsupported.
- Lifecycle confirmation plans are in-process objects and must be recreated from archive replay after restart.
DSH Compatibility
Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.
Security Signals
Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.
package.json declares 0 runtime, 10 development, 0 peer, and 0 optional dependencies.
GitHub reports the repository license as MIT.
A root package.json was captured and can be inspected by the registry.
Public GitHub source metadata is available for this registry snapshot.
Source & Registry Notes
Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.
- Source repository
- mianyoubiaoqing/dsh-Mmem
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- 6bd421b · 2026-09-13
This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.
Repository Activity
- GitHub stars
- 7GitHub stars
- Forks
- 0
- Open issues
- 20
- Last commit
- 2026-08-21
- Last release
- No release detected
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Second-model AI auto-review for DeepSeek Harness approval requests with fail-closed safety and session audit.
View plugin →Browser automation plugin enabling DeepSeek Harness agents to interact with logged-in browser sessions safely via dedicated Agent Windows.
View plugin →Persistent multi-model Agent teams and bounded DAG workflows with GUI controls for DeepSeek Harness.
View plugin →A DeepSeek Harness Web UI sidebar for tracking artifacts and browsing workspace files, with a pop-out tab view.
View plugin →