dsh-knowledge-sqlite
Run the following command in DeepSeek Harness:
dsh plugin install NinjaSln-labs/dsh-knowledge-sqlite
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install NinjaSln-labs/dsh-knowledge-sqlite in your terminal to install the plugin; full source is available at https://github.com/NinjaSln-labs/dsh-knowledge-sqlite
About this plugin
DeepSeek Harness ships with a unicode61 full-text search engine that matches zero Chinese substrings. dsh-knowledge-sqlite replaces that baseline with SQLite FTS5 trigram indexing so CJK substring search actually works, then layers L1 query expansion on top—one lightweight model call per query, cached by normalized query, with a timeoutMs fallback to pure lexical retrieval (degraded: 'lexical' metadata, never a tool error). On the project's hard-query evaluation set, recall jumps from 0% to 21–64% at L1.
The plugin implements the full V1.11 knowledge seam contract: knowledge_write, knowledge_search, knowledge_update, knowledge_list, and knowledge_delete model tools, backed by dedupeKey idempotent upserts, TTL expiry, cursor pagination, and event broadcasts (written / updated / deleted). On the security side it enforces workspace isolation, ask-gated write/delete, authorTier anti-spoofing, and a global writer allowlist so multi-agent setups never leak across boundaries.
If you are building a long-term-memory agent on DeepSeek Harness—especially for Chinese knowledge-base Q&A, project experience capture, or cross-session context continuity—this plugin drops in as a single-file SQLite dependency and plugs a durable, auditable knowledge layer straight into your workflow.
Use Cases
- Cross-session knowledge persistence and retrieval
- Chinese-language knowledge base Q&A and experience recall
- Shared knowledge management with workspace isolation
Best For
- Developers building long-term-memory agents on DeepSeek Harness
- Knowledge-base projects requiring CJK substring retrieval
- Teams needing an auditable and isolated knowledge foundation
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.