Mémoire

Model-authored context pruning for DeepSeek Harness through the official compaction API.

Mémoire de passation auto-entretenue par répertoire de travail et branche git : enregistre les tours, les condense en Markdown concis, et injecte le résultat dans les futures sessions depuis ~/.agent/agent-handoff, compatible octet par octet avec pi-handoff.

Wiki Markdown local en guise de mémoire à long terme pour DeepSeek Harness — portage de llmwiki.

Mémoire de persona persistante à long terme pour DeepSeek Harness : stockage MEMORY.md/USER.md, outils memory + memory_search, injection de contexte par requête, et détection de secrets. Portage dans l'esprit de pi-hermes-memory.

Un plugin DeepSeek Harness qui transforme la synchro git anywhere-claude-mem en pair DSH : récupère la mémoire inter-machines au début de session et pousse les nouveaux prompts/observations à mesure qu'ils arrivent dans le worker claude-mem.

Backend de mémoire vecteur + graphe : outils observe/remember/search/recall vers un serveur memoria local, fusionnant le rappel sémantique HNSW, la recherche par mots-clés FTS5 et les signaux de graphe de connaissances via un classement RRF, avec persistance automatique en fin de tour et isolation par espace de noms par agent.

Passerelle vers une base de connaissances Markdown globale partagée avec la CLI Codex kb.cmd : outils kb_add/kb_search/kb_show/kb_timeline avec un frontmatter compatible octet par octet.

Coffre de mémoire inter-sessions : outils remember / recall / forget, injection de prompt par tour, et un navigateur d'entrées accessible depuis la page de réglages.

A super-simplified LLM-wiki memory plugin: one index document (auto-loaded) + one markdown file per topic (read only when needed) — no dumping everything into the context and burning tokens. Simple and lightweight, painless to install/uninstall, and freely editable however you like.

DSH自记忆插件：结构化记录问题、方案与模型调用，支持加权检索、前提感知排序与外部导入。

DeepSeek Harness 分层长期记忆插件：瞬时/情景/语义三层记忆，自动捕获、AI 提炼、跨会话召回与可视化管理。

DeepSeek Harness (DSH) plugin for local-first automatic Wiki memory and native Skill generation.

Markdown folder long-term memory for DeepSeek Harness: one file per fact, an index line per memory, mounted into every request as a system prompt section

DeepSeek Harness (DSH) plugin — inject context when user input or model output hits configured keywords, with toggleable/configurable rules and model-editable memory.

Industrial-grade 4-tier cognitive memory, red-light guard, and vector semantic engine for DeepSeek Harness (dsh).

Import chat history and memory from other AI agents (Claude Code, Codex CLI, Gemini CLI, Aider, Cline, OpenCode, Hermes Agent, OpenClaw) into DeepSeek Harness

DeepSeek Harness community Bundle for local, evidence-backed Agent Memory

File-based self-evolving memory for DeepSeek Harness v2: compression & extraction over official compaction, daily episodic layer, dream consolidation to semantic files (user/agent/memory/dream + project), catalog disclosure injection, no confidence machin

Notion connector for DeepSeek Harness: search, read, query, create, update, and append Notion pages and databases via six agent tools, with a Web settings page — configured once with an Integration Token.

Write notes and memos to flomo (浮墨笔记) from any agent via the flomo_send tool, configured once with the flomo API URL or API Key.

Auto-supplement QMD semantic search when the model greps/globs a knowledge-base directory: zero-dependency DSH plugin, async injection via next-step, no manual invocation.

DSH ↔ agentmemory session-memory bridge: mirrors session lifecycle into a local agentmemory daemon (REST), exposes memory_recall / memory_remember tools, and injects a per-session memory context window via agent/pre-step.

Tracks the agent's state as Yogācāra mental factors over a durable seed store, and feeds it back into the prompt as first-person inclination.

Bi-temporal memory backed by a SodaMem daemon: recall is contributed to the prompt while each turn is assembled and every closed turn is ingested, with no tool call and no model call on the read path; each recalled fact cites the turn it came from and carries validity dates, so superseded facts stop being retrieved.