dsh-plastic-memory
Run the following command in DeepSeek Harness:
dsh plugin install KKL08/dsh-plastic-memory
Paste the following prompt into your AI chat to install this plugin:
Install it in DeepSeek Harness by running dsh plugin install KKL08/dsh-plastic-memory; the full source is available at https://github.com/KKL08/dsh-plastic-memory .
About this plugin
Agents that run long-term inevitably accumulate stale, contradictory, and redundant memories, leading to cognitive drift, hallucinations, and conflicting decisions. dsh-plastic-memory delivers a full memory system for DeepSeek Harness: it extracts structured memory nodes from conversations, stores them as local Markdown files under five built-in categories (profile, preference, knowledge, project, reference), keeps workspaces physically isolated from one another, and mandates explicit user confirmation before any memory is promoted to the shared global scope.
On the retrieval and governance side, the plugin ships with multi-dimensional full-text search, causal traceability (every memory is anchored to its originating conversation slice), a 0-100 health score, and deep whole-library scans that detect API key leaks, semantic conflicts, redundancy, and orphaned references. A decayDays freshness mechanism lets expired information retire gracefully, while three recall injection modes - core (full-text resident), search (index on demand), and passive (silent index) - balance context quality against token budget.
If you want your agent to maintain precise, controllable long-term memory without the output-quality degradation caused by uncontrolled memory bloat, this is a complete memory governance solution in the DeepSeek Harness ecosystem built around the ideas of plasticity and lifecycle.
Screenshots
Use Cases
- Periodic governance and graceful decay when a long-running agent accumulates contradictory, redundant, or stale memory entries
- Isolating per-workspace memories while sharing global preferences across teams to prevent cross-contamination
- Tracing memory provenance to verify original intent and context when an agent produces hallucinations or conflicting decisions
Best For
- Developers building production-grade agents on DeepSeek Harness
- Teams managing multi-project memory isolation and global promotion workflows
- Engineering practitioners focused on long-term memory quality, preventing cognitive drift and token waste
Related Plugins
Memory layer for coding agents that indexes local session history and auto-recalls relevant context before edits, commands, and failures, with no manual search needed.
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.