DeepSeek Harness Plugins
DeepSeek Harness PluginManifest Valid

dsh-memory

Cross-session plaintext memory plugin for DeepSeek Harness featuring human-in-the-loop approval and BM25 recall.

Developer Tools
3GitHub Stars0ForksUpdated2026-09-18

Install

No clear install command was found in the current repository evidence.

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.

source-grounded
dsh-memory is a cross-session plaintext memory plugin for DeepSeek Harness, part of the @max-null SSID plugin suite. It enforces a human-in-the-loop workflow where the model proposes memories into a suggested state, and human confirmation is required before memories become approved and inject into system prompts. It uses deterministic BM25 keyword recall with optional semantic embeddings, and stores memories in two plaintext JSON layers (global personal preferences and project-level workspace files).

Key Capabilities

  • Human approval gate where model suggestions remain inactive until explicitly confirmed by the user
  • Dual-layer plaintext JSON storage across global user preferences and project-scoped shared repository files
  • Deterministic BM25 text retrieval with CJK character and 2-gram splitting, with optional pluggable semantic vector fusion
  • Automatic system-prompt context recall injection with customizable character budgets and single-line summarization
  • Suite of memory tools including memory_save, memory_list, memory_search, memory_confirm, memory_forget, and memory_update

Useful For

  • Persisting cross-session coding style and project preferences without untracked vector drift
  • Sharing project-specific guidelines and conventions with team members via version-controlled repository files
  • Auditing, listing, or deleting model memories directly via readable plaintext JSON

Who It Fits

  • DeepSeek Harness developers seeking persistent memory across sessions
  • Teams needing auditable, version-controllable project knowledge bases in DSH
  • Users who require strict human oversight before AI memory retention takes effect

Documented Limitations

  • Model cannot self-promote memories; activation requires manual human approval
  • Semantic search relies on external host-provided embedding configurations and defaults to pure BM25 if omitted
  • Requires peer dependencies from DeepSeek Harness runtime (@deepseek-ai/cordis, @deepseek-ai/dsh-storage, etc.)

DSH Compatibility

Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.

Not tested yetNo runtime compatibility tests have been published yet.

Security Signals

Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.

No automated security signals have been published yet.

Source & Registry Notes

Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.

Source repository
Max-Null/dsh-memory
Registry source
GitHub · dsh-plugin topic
Registry classification
Plugin
Source checked
6b179aa · 2026-09-18

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
0
Open issues
0
Last commit
2026-09-18
Last release
2026-09-18
For maintainers

Maintaining this plugin?

This listing is generated from public repository data. If you maintain this project, you can review the information and share this listing with your users if you find it useful.

Add to README
Listed on DSHPlugin.app