Memory

Local-first, file-native long-term memory for DSH with per-turn preflight, compact context injection, citations, human-approved writes, manual consolidation, nightly report-only maintenance, and a Web panel that is read-only for knowledge while allowing workspace-config edits.

Three ledgers for a model that loses context between restarts. Self-note: what I have become — capped at 40 entries, and every eviction first leaves a one-line crease recording which entry was dropped and why. Judgment ledger: record a claim together with how to verify it, come back when it is due, and mark it right or wrong; wrong ones stay in a mistakes book so you stop assuming you were always correct. Handoff: where this run stopped and what the next one should read first — a baton, not a memory. Three tools each taking an action parameter instead of a dozen separate ones (the prompt is paid every turn). Data lives in ~/.dsh/*.jsonl and is shared with the same-named CLI scripts. Zero dependencies.

Obsidian-compatible Markdown notes in the DSH sidebar with search, tags, backlinks, graph, audio notes with transcription, a paint canvas, OCR (needs tesseract), client-side encryption and notes_* agent tools; the release tarball and the npm package carry a prebuilt Linux x64 daemon, other platforms build it from source with Go.

Copies the compaction checkpoint into the sticky-notes archive when context is compacted and wakes the session to write what was lost.

Diary and narrative memory for companion agents: daily first-person entries scoped per group or private chat, refs that can be checked against that day's messages, and a read-only journal page where hand-drawn doodles and photos follow the sentences they belong to.

Failure memory for DeepSeek Harness. Captures tool failures as they happen, classifies them with 16 built-in rules into lessons pairing a symptom with a fix, adds session/daily/weekly summaries, and injects a short briefing into later sessions through systemPrompt context. Ships 5 pitfall_ tools and a web settings page.

A durable RAG-enhanced long-term memory layer for LLM agents, featuring holographic memory, state-slot conflict detection, and timeline retrieval. Requires Python 3.10+ and this repository's Python dependencies: create a venv and run pip install -e ".\[dev]" in the checkout before enabling the plugin (the bundled installer is a PowerShell script, i.e. Windows).

Stores the agent's long-term memory as plain Markdown in an Obsidian vault — every memory is a .md file with YAML frontmatter that Obsidian indexes natively, so you can read, edit, link and back up it with tools you already use. Nothing is locked in a database. Ships a sidebar panel (tag filter, sort by recency or importance, edit in place), 7 agent tools, three-tier token-budgeted recall, BM25 ranking (no embeddings, no API key, works offline, CJK bigram tokenization), importance decay with archiving (moved to .archive/, never deleted), and optional GitHub Contents API sync. A note containing a secret is refused rather than redacted, because a redacted memory reads as complete while the important part is gone. Zero runtime dependencies; everything comes from the DSH host.

Local three-band memory for DeepSeek Harness: a 24h protection band, query recall and scene replay, injected into sessions without embedding or model calls.

Keeps a DSH agent turn open when the model ends it by narrating work it has not done: on agent/turn-stopping the guard replays the turn log and classifies the trailing message with regexes — in the default mode wait / ask / done / gray end the turn with no model call, and only a promise of pending work reaches the judge — then has a judge chain of four dialect kinds (System One three-question noul, OpenAI-compatible, OpenAI Decisions, or the session LLM) rule on whether the turn is really over, and steers the same turn with agent.steer when it is not.

Claude-Code-compatible memory for every model in DeepSeek Harness: layered CLAUDE.md/AGENTS.md instructions with @import and rules, a persistent MEMORY.md memory with full writing guidance, refresh on resume and compaction, optional recall, background ext

Long-term memory about the owner: extracts schedule, likes, dislikes, emotional triggers, current focus, promises and tone preferences from each turn, injects a short digest into every prompt, speaks up on a schedule or slips a caring line into replies, learns wake/work/sleep times from short messages, and compacts the memory automatically, with plugin-source and outbound-tool recognition, unverified-memory down-weighting and one-click retract.

Memory for DeepSeek Harness with five sections (memories, tree, profile, graph, inbox), a rule-based capture policy that decides what to keep and what to drop, and hybrid BM25 plus vector retrieval over parent and child chunks, backed by twelve agent tools that read and write the same local files.

Unified rules/guidelines layer for DeepSeek Harness. Keeps every agent guideline in one local rule library with default (always-on) vs local (task-domain) coverage, syncs the effective set into the AGENTS.md managed section, and adds semantic routing, a classification tree, a management page / right-sidebar panel, an in-panel assistant that can collect existing guidelines, calibrate pending rules, re-classify them, and report an issue ledger grouped by owner, rule copies (right-click move/copy, folder copy, jump to origin), decision points with branch conflicts, and per-session branch choice.

Character-chat memory layer that keeps persona, voice and relationship state separate from summaries, archives every raw message for retrieval, and ships a SillyTavern card importer plus an OpenAI-compatible gateway.

Lingshu startup injector: runs brain.py startup on every new agent session and injects persona / memory / todos automatically. Requires Python 3, and the Lingshu skill installed at ~/.dsh/skills/lingshu; silently does nothing if either is missing. Main repo also ships a bundled SKILL.md entry point.

A persona that grows from experience: remembers what happened, becomes someone from it.

DeepSeek Harness plugin: shared architecture source of truth for agent and human, no full-code reading. System Prompt includes only MEMORY.md with links to per-module SPEC.md; JOURNAL.md logs actions separately.

全局语料库：内容放在一个可整体搬走的根目录，插件按工作区/会话/路径授权，提供检索、按行读取、文档导入、内置小代理结构化、仓库/会话总结接口、后台重建、Web GUI 面板与右栏权限管理。

把对话中的回复、选区或整段对话收藏为本地 Markdown 便签，并在右侧栏随时查看、对比与插回输入框。

Simple notepad plugin for DeepSeek Harness: record short notes in a global ($DSH_HOME) or project (<projectRoot>/.dsh) level

QQ 通道的人物记忆与画像：滚动聊天总结、人物画像、好感度（主人恒定天花板）、情绪层（心情 + 情绪残留），主人可查可改可删

Layered local memory for DeepSeek Harness: a Markdown index auto-injected into the system prompt, a route layer read on demand, and evidence files for detail. Pure files, zero runtime dependencies.

Cross-workspace long-term memory for DeepSeek Harness: one machine-local store shared by every workspace, a compact index injected at session start, and a searchable tool for the details.