dsh-local-memory: A Cross-Project Long-Term Memory Plugin for DeepSeek Harness
The dsh-local-memory plugin is designed to fill the management gap of the Hindsight memory component in DeepSeek Harness (DSH) across multi-project scenarios. It preserves the “one database per project” isolation model while providing cross-database aggregate views and diagnostic tools, addressing the lack of a holistic perspective and the difficulty of diagnosing unused projects. Its core features include scanning project status, cross-database queries, bulk memory creation (with support for seeding from Git history), and a graphical management interface. Installation depends on the Hindsight service and requires local daemon mode and an LLM key. The plugin queries via aggregate views, displays results grouped by project, and supports specifying the query scope through parameters to optimize performance.
Read Moredsh-local-memory: DeepSeek Harness Local Cross-Session Memory Plugin
dsh-local-memory is a local memory persistence plugin designed for DeepSeek Harness. By automatically capturing conversation summaries, injecting short-term memory (default: 2 days), and providing keyword-based retrieval tools, it addresses memory loss caused by the model's limited context window. Implemented in pure JS with no external library dependencies, the plugin supports actively retrieving historical information when conversations exceed the context window and can share a data directory with opencode's memsearch.
Read More