dsh-dream
Run the following command in DeepSeek Harness:
dsh plugin install STARDUSTLC666/dsh-dream
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install STARDUSTLC666/dsh-dream in the DeepSeek Harness terminal to install; the source repository is at https://github.com/STARDUSTLC666/dsh-dream. Restart the Web service after installation to activate.
About this plugin
Your agent produces dozens of conversations every day, yet every session evaporates the moment it ends. How a problem was actually solved, which tool call went sideways, which user preferences surfaced again and again - these valuable signals remain buried in raw logs, never distilled, replayed, or consolidated. dsh-dream borrows the mechanism of human sleep, where the brain replays the day and cements memory, and gives the agent the same ability. It reads past sessions, extracts dream material, reflects, and writes a permanent dream journal so that experience actually sticks.
The core workflow forms a tight loop. dream_digest replays recent sessions, automatically skipping sub-agent noise and restoring streamed text from packed frames. dream_save persists reflections, one to five lessons, and a mood into a JSONL journal. dream_recall retrieves any past dream by keyword. The critical step is dream_bridge: it idempotently merges the highest-frequency lessons into AGENTS.md inside a tagged block, so repeated runs only refresh the block without corrupting the file - recurring dreams become genuine long-term memory. All secrets and tokens are auto-masked before ingestion, and the bundled dream-protocol skill teaches the agent when to dream and how to distill patterns rather than narrate transcripts.
It is built for users who treat the agent as a long-term collaborator and want it to accumulate experience over time without standing up a heavyweight RAG pipeline. It modifies no session files, makes no network requests, and slots into the existing DeepSeek Harness workflow with a single cordis patch - giving the agent a quiet corner to dream, reflect, and slowly grow.
Use Cases
- Auto-extract lessons from past sessions into a permanent dream journal
- Idempotently merge high-frequency lessons into AGENTS.md as long-term agent memory
- Recall any past dream by keyword for quick experience and decision reference
Best For
- Long-term users who want the agent to accumulate experience rather than start from scratch every time
- Light-setup users who prefer to avoid a heavyweight RAG pipeline
- Privacy-conscious users who need automatic secret masking before ingestion
Related Plugins
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.
An opinionated, zero-infrastructure file-based memory protocol: human-readable markdown files plus a bounded always-loaded index, curation discipline, and no database or embeddings.