Memory and knowledge · DeepSeek Harness plugins — Page 23

Self-hosted mem0 memory operations for dsh: agent tools (mem0_add / mem0_search / mem0_get / mem0_update / mem0_delete / mem0_history / mem0_reset / mem0_status) against your own mem0 REST server (new OSS build, X-API-Key auth, no /v1 prefix). Hot-pluggable — mounted via `dsh plugin --profile web add link:<path>`, no dsh source changes.

Long-term memory for DeepSeek Harness: multi-channel retrieval and an evolving memory store.

Project-scoped memory system for DeepSeek Harness: persistent key-value memories per scope with set/get/list/delete/search tools, JSON-file storage and client RPC. · DSH 项目记忆系统：按作用域持久化的键值记忆，含增删查列搜索工具与 RPC。

Durable cross-session memory for DeepSeek Harness: local SQLite FTS5 store, three model-facing tools, and a recall prompt section — no embedding service and no sidecar

Persistent three-layer memory for DeepSeek Harness: host-side cross-workspace store, per-turn auto-consolidation, and memory tools for the agent. 持久三层记忆：宿主跨工作区存储 + 每回合自动沉淀 + agent 记忆工具。

Cross-session persistent memory for the DeepSeek Harness: a model-facing memory tool, an always-on index section, and a bundled maintenance skill.

Settings → Memory: Noema long-term recall, search, remember, and import

Memory retrieval plugin for DeepSeek Harness (dsh): wiki double-link memory graph with 9 tools — search (hybrid vector + fulltext), read, list, link-scan, remember, forget, compile, health-check, and a resident retriever subagent

Cross-session memory for DeepSeek Harness (dsh): background Auto-Memory, per-workspace daily notes, Dream consolidation, memory/session search, and optional vector search

Evolving long-term memory for the DeepSeek Harness: durable, decay-reinforced memories with recall, consolidation, and pruning.

Memory Fortress for DeepSeek Harness: 5-track layered memory, self-evolution, knowledge graph (SQLite+FTS5), status pill. 本地记忆堡垒：五轨记忆·自我进化·知识图谱·状态灯。

为 AI Agent 提供短期记忆缓存、长期知识沉淀、语义检索与持久化存储的统一记忆管理层

把 WorkBuddy 的文件式记忆系统移植进 DeepSeek Harness：用户级 + 工作区的人类可读 Markdown 记忆（~/.deepseek-harness/MEMORY.md 与 <cwd>/.deepseek-harness/memory/）。

Cloud-backed cross-session memory for DeepSeek Harness: structured entries + vector retrieval persisted to S3-compatible object storage (AWS S3 / MinIO / R2 / OSS), with local cache projection, approval-gated writes and auditable trails

Solve the AI goldfish brain: a personal memory layer for DeepSeek Harness — preferences, project conventions, workflows, and error lessons, stored locally as auditable, evidence-backed Markdown/JSON. 解决 AI 金鱼脑：偏好、项目约定、工作方式与纠错教训的本地可审计记忆层。

Memory Standard Protocol (mm) for the DeepSeek Harness: layered MEMORY.md + deterministic budgets + frozen snapshots + cross-agent markdown/URI interop + session-log digestion. Ships as a dsh bundle (tools: mem_read, mem_write, mem_search, mem_budget, mem_digest) and a standalone CLI.

Local-first persistent memory infrastructure for DeepSeek Harness: bounded hot-memory bootstrap, explainable cold recall (exact + Chinese BM25), lease-lock transactional writes, read-only governance, and trajectory review. Pure Python + Markdown, zero external dependencies.

Four-layer long-term memory for DeepSeek Harness (L0 conversation → L1 atoms → L2 scenes → L3 persona) with auto-recall and auto-capture — adapted from TencentCloud/TencentDB-Agent-Memory

DeepSeek Harness 记忆分诊插件：自动判断记忆保留/归档，并自动路由到项目桶或总体桶（规则优先 + LLM 兜底）

DeepSeek Harness plugin: inject MemVault core memory blocks into the system prompt, extract finished turns as one windowed pass, and show, edit, trace, review and replay all of it in a memory panel

MindsEye: intent-routed vision, structured evidence, and exact cache for DeepSeek Harness

MiOpIIk plugin suite for DeepSeek Harness: one-command install of all seven plugins plus the miopiik four-layer workflow preset initializer. MiOpIIk 套件包：一条命令装齐 7 插件并初始化四层工作流 preset

Non-destructive DSH seam capability probe writing a two-evidence-level manifest (present vs invoked-ok). DSH seam 能力探测，输出现场/实调双证据能力清单

MiOpIIk 里程碑自动检查点：根会话每轮关闭时自动把一行恢复点追加到 .dsh/memory/checkpoints.md，由 agent/turn-stopping 结构事件驱动，不依赖会话自觉