Install
$ dsh plugin --profile web add dsh-layered-memoryPlugin 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
- Automatic L0-L3 conversation capture and background distillation pipeline
- Pre-step synthetic message recall injection and system prompt persona integration
- Hybrid retrieval combining BM25 full-text search with optional local (ONNX CPU) or remote vector embeddings
- Session-level memory scope gears (Auto, Work, Chat, Read-Only, Paused)
- Built-in agent memory tools: memory_search, conversation_search, and memory_read_scene
- Interactive Memory Workbench UI in settings with cost analytics and asset maintenance
Useful For
- Maintaining user preferences, project background, and team methodologies across distinct sessions
- Reducing token consumption and exploratory steps in long, multi-step engineering tasks
- Separating personal and work contexts automatically without manual tagging
- Enabling local-first, privacy-conscious semantic memory indexing without external vector DB dependencies
Who It Fits
- DeepSeek Harness web profile users seeking persistent cross-session memory
- Developers working on multi-session software projects requiring continuous context
- Power users desiring fine-grained control over agent recall scopes and LLM distillation budgets
Documented Limitations
- Requires DeepSeek Harness host version >= 0.1.2-alpha.1 and Node.js >= 22.16.0
- Vector embedding capabilities require either configuring remote embedding endpoints or downloading local ONNX models into the data directory
- Distillation consumes background LLM tokens using the host ctx.llm or configured layer routes
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 4 runtime, 15 development, 8 peer, and 0 optional dependencies.
package.json declares DSH bundle metadata.
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
- JunNanLYS/dsh-layered-memory
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- 69028fb · 2026-10-12
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
- 20GitHub stars
- Forks
- 5
- Open issues
- 9
- Last commit
- 2026-09-07
- Last release
- 2026-09-07
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Lingshu long-term auditable memory system and cognitive graph for DeepSeek Harness.
View plugin →Knowledge graph memory and context engine for DeepSeek Harness with cross-session recall.
View plugin →Composable, view-based layered memory plugin bundle for DeepSeek Harness with pluggable sources and providers.
View plugin →Comprehensive long-term memory, self-evolution, skills, task management, and multi-agent dispatch bundle for DeepSeek Harness.
View plugin →