dsh-prime-memory
Hierarchical L0–L3 distilled memory plugin for DeepSeek Harness with hybrid retrieval and automated context injection.
Install
$ dsh plugin --profile web add dsh-prime-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 to L3 hierarchical memory capture, extraction, consolidation, and distillation
- Pre-step synthetic message injection into context with deduplication and recency decay weighting
- Hybrid search combining BM25 full-text indexing (via jieba) and vector search (remote or local ONNX)
- Session-level memory mode switcher (daily, work, auto, off, write-only)
- Comprehensive memory tools for agents including search, scene reading, conflict resolution, and graph traversal
- Cost telemetry tracking per-model token consumption and execution latency
Useful For
- Long-running tasks and agent workflows needing continuous context retention across multiple sessions
- Maintaining user persona and team methodologies across varied interaction tracks
- Reducing repeated exploratory tool calls in multi-step coding or research tasks
- Selective memory capture for privacy-sensitive or benchmark testing sessions
Who It Fits
- DeepSeek Harness developers building stateful or long-horizon agent workflows
- Power users requiring cross-session persistence of preferences and project context
- Teams evaluating agent memory benchmarks and token efficiency
Documented Limitations
- Requires Node >= 22.16.0 and specific DSH engine versions (0.1.x to 0.2.x rc series)
- Local embedding features require downloading ONNX models and transformers.js into the data directory
- Upstream provider TTFT timeouts may interrupt distillation if fallbacks or layer-specific routes are not configured
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 5 runtime, 23 development, 10 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
- drscrewdriver/dsh-prime-memory
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- b9c9682 · 2026-10-26
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
- 15GitHub stars
- Forks
- 3
- Open issues
- 6
- Last commit
- 2026-10-24
- Last release
- 2026-10-22
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Persistent semantic memory layer for DeepSeek Harness backed by Markdown and Milvus.
View plugin →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 bundle for DeepSeek Harness with pluggable providers and strategies.
View plugin →