dsh-session-archive-plugin
Run the following command in DeepSeek Harness:
dsh plugin install honyKing/dsh-session-archive-plugin
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install honyKing/dsh-session-archive-plugin to install the plugin from https://github.com/honyKing/dsh-session-archive-plugin.
About this plugin
DSH ships with compaction-basic, which auto-compresses context once usage hits 80%, but the compressed logs remain as multi-frame zstd binaries—hard to read and impossible to search. When a key conclusion scrolls out of view, users are stuck flipping through raw binary logs. This plugin fills exactly that gap.
It exposes two model-callable tools: archive_session decodes the current session log into plain-text JSONL plus a Markdown summary saved under session_archives, with an idempotent guard that skips re-archiving when less than the configured threshold of new data has accumulated. search_archive first scans Markdown summaries across workspaces; if nothing matches it transparently decodes the most recent raw logs for a direct scan, supporting multi-word AND, limit, and deep parameters so history is always reachable. A bundled skill also nudges the model to proactively search archives after compression, closing the read-write loop.
Best for DSH users who run long, multi-turn reasoning sessions and need to retrieve specific inferences or decisions after compaction. If your workflow involves accumulating dense context and then hunting for a single detail, this plugin keeps every fact one search away, before or after compression.
Use Cases
- Retrieving key conclusions after context compression
- Searching specific inferences or decisions across sessions
- Converting binary session logs into readable Markdown archives
Best For
- Developers running long multi-turn DSH sessions
- AI engineers tracing reasoning chains across conversations
- Power users who value full context retention
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.