dsh-hindsight
Run the following command in DeepSeek Harness:
dsh plugin install Ryu6Zero/dsh-hindsight
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install Ryu6Zero/dsh-hindsight in the DeepSeek Harness terminal to install this plugin (source: https://github.com/Ryu6Zero/dsh-hindsight), then restart the target Web or Headless profile to activate all tools and slash commands.
About this plugin
Every DSH Agent session starts from scratch—that is the default, and it is also the pain. Existing memory solutions tend to arrive as a fully orchestrated three-layer stack: heavy to deploy, hard to debug. If your team already runs Hindsight, the vector memory engine with entity resolution, knowledge graphs, and multi-strategy retrieval, the only thing missing is a thin, self-contained bridge. dsh-hindsight is exactly that: a zero-dependency HTTP client that plugs DSH straight into the Hindsight REST API. No bundled storage, no orchestration layer, just one installable package.
Once connected, the agent gains a full cross-session memory toolkit. Slash commands and six model tools let it recall past decisions, persist new facts, and forget stale entries. The hindsight_related tool walks the Hindsight entity graph up to five BFS hops, threading connected decisions and entities into working context. With autoRemember enabled, the model proactively surfaces durable preferences, constraints, and decisions during conversation and proposes saving them after a quick confirmation. hindsight_condense batches two to ten facts per write with conservative text-normalisation dedup. Identical queries served from a 60-second recall cache, invalidated the moment any write lands. Every critical setting—endpoint, token, bankId, timeouts, autoRemember—is hot-reloadable, so changes take effect without a profile restart.
This plugin is built for teams that already operate Hindsight as their shared vector-memory backend and want DSH agents to carry real long-term memory without pulling in a heavyweight orchestration framework. One install, one slash command, six tools. Done.
Use Cases
- Let DSH agents recall user preferences and key project decisions across multiple sessions
- Walk the entity-relationship graph from a single memory to surface linked constraints and context
- Batch-submit new facts with automatic deduplication against recent memories to prevent pollution
Best For
- Teams already running Hindsight who want DSH agents to gain long-term memory immediately
- Developers who prefer a lightweight single-package integration over heavy orchestration stacks
- Platform admins who need cross-bank read-only search with strict write isolation
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.