dsh-layered-memory
Cross-session long-term memory plugin for DeepSeek Harness.
Install
$ dsh plugin --profile web add github:DDDFXYqiming/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
- Injects an L1 memory index into system-prompt context on each turn.
- Stores L2 environment facts and L3 task experience/SOPs.
- Provides memory listing, reading, writing, updating, archiving, rollback, and source-event expansion tools.
- Requires evidence when writing memory through memory_write.
- Supports pending candidates distilled from successful tool calls and acceptance into formal memory.
- Tracks provenance metadata including source session, sequence references, timestamps, and evidence.
- Supports namespace isolation, with automatic workspace and Git-branch-based namespaces.
- Provides maintenance for deduplication, index compression, statistics, and pending-candidate merging.
Useful For
- Retain verified workspace facts across separate DSH sessions.
- Reuse validated procedures and task experience through stored SOPs.
- Review and accept automatically generated memory candidates.
- Audit a stored memory item's source session and events.
- Archive or roll back superseded memory while retaining historical files.
Who It Fits
- DeepSeek Harness users who need persistent cross-session memory.
- Agent workflow builders maintaining verified facts and reusable procedures.
- Developers using DSH profiles, Cordis configuration, and agent skills.
Documented Limitations
- Requires Node.js 22.19 or later.
- Requires compatible peer dependencies including @deepseek-ai/cordis version 4 release candidates below version 5.
- memory_write requires evidence; unverified information cannot be written through that tool.
- The L1 index is capped by maxIndexLines, defaulting to 30 logical lines.
- Automatic maintenance runs every 20 turns by default unless configured otherwise.
- Archived memories are hidden from the L1 index and memory_read, though retained in archive files.
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.
Source & Registry Notes
Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.
- Source repository
- DDDFXYqiming/dsh-layered-memory
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- 3729dfe · 2026-09-19
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
- 3GitHub stars
- Forks
- 1
- Open issues
- 0
- Last commit
- 2026-09-18
- Last release
- No release detected
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Comprehensive long-term memory, self-evolution, skills, task management, and multi-agent dispatch bundle for DeepSeek Harness.
View plugin →Composable three-tier memory control plane for DeepSeek Harness with runtime context, searchable documents, and pluggable long-term memory.
View plugin →A standalone, local persistent-memory system for AI agents, with DeepSeek Harness support provided through the separate dsh-mnemon plugin.
View plugin →Local, fast cross-agent memory and session indexing system with DeepSeek Harness integration.
View plugin →