← DeepSeek Harness Plugins
DeepSeek Harness PluginManifest Valid

feiyu-memory

Cross-session persona and episodic diary memory plugin for DeepSeek Harness Desktop.

Developer Tools
4GitHub Stars1ForksUpdated2026-10-21

Install

$ dsh plugin --profile desktop add github:AirPineApplea/feiyu-memory

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
feiyu-memory is a cross-session persona and memory management plugin for DeepSeek Harness Desktop. It utilizes local gate checks to evaluate conversation rounds without token overhead, selectively summarizing noteworthy interactions into first-person diary logs and qualitative impressions. Prior to new conversation rounds, it injects a capped memory block (~2000 tokens) into system prompts using layered sections (persona core, self-judgments, and episodic diary entries) backed by plain JSONL and BM25-lite bigram retrieval.

Key Capabilities

  • Zero-token local turn-stopping gate to evaluate whether interactions warrant archival
  • Layered memory representation separating persona core, AI self-notes, and episodic recollections
  • BM25-lite bigram inverted index for local keyword-based associative recall
  • Decoupled token budget capping memory injections at a default of 2000 tokens
  • Automated local privacy redaction for API keys, sensitive credentials, and third-party entity names
  • Desktop sidebar client panel to review, delete, or export memory packages (.dshpack)
  • Plain-file storage in ~/.dsh/feiyu-diary/ with zero external runtime dependencies

Useful For

  • Retaining consistent AI persona and rapport across multiple chat sessions in DeepSeek Harness Desktop
  • Preventing repetitive instructions and conversational context loss without ballooning context window token costs
  • Exporting and transferring personalized AI memory archives between machines or accounts

Who It Fits

  • DeepSeek Harness Desktop users seeking persistent conversational memory and companion persona depth
  • Community experimenters exploring lightweight local-first LLM memory architectures

Documented Limitations

  • Intended as a casual experiment and hobby project, explicitly not audited or suitable for sensitive enterprise or workplace production data
  • Requires DeepSeek Harness Desktop to be fully terminated before running the profile install command
  • Uses bigram keyword scoring (BM25-lite) rather than semantic vector embeddings for memory retrieval

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.

Dependency Counts

package.json declares 0 runtime, 0 development, 1 peer, and 0 optional dependencies.

info
Dsh Bundle Declared

package.json declares DSH bundle metadata.

info
License Declared

GitHub reports the repository license as NOASSERTION.

info
Package Manifest Available

A root package.json was captured and can be inspected by the registry.

info
Source Available

Public GitHub source metadata is available for this registry snapshot.

info

Source & Registry Notes

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

Source repository
AirPineApplea/feiyu-memory
Registry source
GitHub · dsh-plugin topic
Registry classification
Plugin
Source checked
3f0f080 · 2026-10-21

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
4GitHub stars
Forks
1
Open issues
6
Last commit
2026-10-21
Last release
2026-10-15
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