dsh-evolve
Self-evolving memory and skill lifecycle engine with deterministic zero-token recall, tiered approval, and anti-bloat pruning for DeepSeek Harness.
Install
$ dsh plugin --profile web add github:chenzheshushi-commits/dsh-evolvePlugin 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
- Durable cross-session structured memory (facts, preferences, decisions, lessons, todos) stored in JSON with readable Markdown mirrors
- Deterministic zero-token hybrid recall fusing bigram-Jaccard similarity and SQLite FTS5 BM25 via RRF
- Tiered approval gate to automatically confirm reversible, user-anchored memories while queuing uncertain writes for review
- Skill crystallization from high-value lessons into versioned, editable SKILL.md documents with pre-operation backups and rollback
- Anti-bloat convergence and controlled pruning (soft-delete, memory budget enforcement, coldness scoring, and duplicate skill merging)
- Isolated background turn review without interrupting main loop context or prompt cache
- Web settings panel integration for memory approval queues, controlled two-stage pruning, and lifecycle configuration
Useful For
- Maintaining long-term user preferences and project-level context across multiple DeepSeek Harness sessions
- Automatically extracting reusable procedural workflows into version-controlled skills based on repetitive agent tasks
- Preventing memory and skill bloat through deterministic deduplication, power-law coldness ranking, and safe archiving
Who It Fits
- Developers and power users building persistent workflows with DeepSeek Harness
- Users needing deterministic, local-first agent memory without relying on vector database APIs or token-heavy embedding models
Documented Limitations
- Requires Node.js >= 22.5.0 due to built-in node:sqlite FTS5 full-text search requirements
- Requires external git and tar binaries on PATH for optional automated memory checkpoints and skill backup rollbacks
- Primary testing performed on Linux and macOS; Windows is untested
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
- chenzheshushi-commits/dsh-evolve
- Registry source
- GitHub · dsh-plugin topic
- Registry classification
- Plugin
- Source checked
- c2f2589 · 2026-09-21
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
- 8GitHub stars
- Forks
- 0
- Open issues
- 0
- Last commit
- 2026-09-20
- Last release
- 2026-09-20
Related DSH Plugins
Ranked by overlapping capabilities, use cases, plugin type, categories, and DSH profile.
Bundle of 88 open-source Agent Skills for research, business intelligence, marketing, and SEO in DeepSeek Harness.
View plugin →Long-term spatiotemporal memory, zero-LLM whitebox reasoning, and knowledge flywheel integration for DeepSeek Harness via AEIS.
View plugin →Composable three-tier memory control plane for DeepSeek Harness with runtime context, searchable documents, and pluggable long-term memory.
View plugin →Traceable, searchable, cross-session knowledge graph memory engine for DeepSeek Harness.
View plugin →