Phant0Meow/dsh-meow-memory
Project-scoped cross-session memory: seven-layer SQLite store (soul/user/project/fact/lesson/rules/topic) on node:sqlite, cache-friendly first-message injection plus per-message keyword hits, memory_remember/search/find_similar/read/update/project tools, and idle-window consolidation ("dream") with peak-hour suppression. Prompt texts ship as per-language data files — the framework supports multilingual users.
Install
via npm
Resolves a published tarball rather than cloning the repository, and installs without any extra setup. Swap `web` for your profile name if you run another one.
via GitHub · npm package
Installing from GitHub runs the project's build script, which pnpm blocks until you allowlist it — run the command once and pnpm prints the exact key to add under `allowBuilds` in ~/.dsh/profiles/web/pnpm-workspace.yaml.
The desktop app keeps its own profile, desktop, and only the dsh command it ships may change it — the npm-installed dsh refuses that profile. Install the command from the app menu (Manage dsh Command…), open Desktop once, quit it fully, run the line below, then reopen Desktop.
via npm
Resolves a published tarball rather than cloning the repository.
via GitHub · npm package
Installing from GitHub runs the project's build script, which pnpm blocks until you allowlist it — run the command once and pnpm prints the exact key to add under `allowBuilds` in ~/.dsh/profiles/desktop/pnpm-workspace.yaml.
Our install check ran in a web profile on npm's dsh, not in the desktop app.
What happened when we ran it
Installed cleanly when we ran it
Every command here is run in a throwaway container against a clean profile, and the result is whatever the harness recorded — not a guess from the source. Last run 8d agoon DeepSeek Harness 0.2.2-rc.1.
Show it in your README
For maintainers: the badge serves this listing's latest sandbox verdict, so a re-run updates it on its own — and it links readers to the full result here.
Due diligence
Security
No build script was blocked in our run.
Source: our sandbox install on DeepSeek Harness 0.2.2-rc.1, 8d ago.
package.json declares no preinstall, install or postinstall script.
Source: package.json on GitHub, read at our last sync. prepare is not recorded; it runs only on a GitHub install, and shows up under build scripts when pnpm blocks it.
Neither shell commands nor credentials.
Source: a keyword match on the README — a hint about what the plugin touches, not an analysis of its code.
Not an audit. These are the record of one install and two text scans; nobody has read this plugin's code on your behalf. Plugins run with your agent's permissions.
What it is — DeepSeek Harness plugin supporting seven-layer SQLite cross-session memory storage along with keyword hit injection.
Who it is for — When you use DeepSeek Harness to manage cross-project sessions and need to retain user information preferences, facts, lessons, and topic rules, this plugin is suitable for installation. If your work is limited to a single independent dialogue session without needing cross-session information accumulation, you do not need to install this plugin.
Watch out — Installation with pnpm may fail due to blocked build script and requires profile configuration to allow the build script; sandbox test passed with successful registration in new profile. No other runtime compatibility issues found.
The verdict — I would install it because it automatically re-injects long-term memory blocks after session compression and supports window-based dream consolidation to suppress peak-hour interference.
Generated by grok-4.6, and a starting point rather than a verdict. Where it says a plugin installs or does not, that is from a real run in a clean profile — everything else is read off the repository. Trust the source over this.
dsh-meow-memory is a DeepSeek Harness plugin that stores and retrieves project-scoped memory across sessions using a seven-layer SQLite database. The layers are soul, user, project, fact, lesson, topic and rules, stored with node:sqlite under the configured projectDir, which defaults to .dsh-meow relative to the workspace.
DeepSeek Harness receives a separate memory snapshot before the first user message, containing long-term soul, user and design-principle records without rewriting the user prompt. From the second message, keyword matching injects up to hitTopK relevant fact, lesson, rules and topic records from the global scope and the current project. The plugin exposes memory_remember, memory_search, memory_project, memory_find_similar, memory_read, memory_update, memory_dream and memory_home, plus the /dream command. After a three-hour idle window, dream can consolidate memories unless the configured peak hours suppress it.
This suits agents that need durable context across work sessions and projects, particularly where explicit memory tools and automatic consolidation are useful. It is a poor fit when workspace-local memory files are undesirable or when retaining project and user context across sessions is not needed. The README documents compatibility with DeepSeek Harness 0.1.5 and use of node:sqlite without native modules; installation does not add a separate runtime dependency.
DeepSeek Harness exposes the plugin settings in its settings page and accepts them through a profile patch or cordis.patch.yml. The main options are:
| Key | Default or documented value |
|---|---|
enabled |
true |
projectDir |
.dsh-meow, relative to the workspace |
promptLang |
zh |
hitTopK |
2 |
dream.enabled |
true |
dream.idleMinutes |
180 |
Additional documented fields include titleMax, autoMigrate, homeDir, disabledWorkspaces, backup and delegate. promptLang selects the prompt and tool-description language from zh, en or pt-br; an unrecognised directory name silently falls back to the Chinese language pack. The interface language follows DeepSeek Harness settings independently.
DeepSeek Harness loads these tools with the plugin:
memory_remember writes content with required project, keywords and importance values, merging duplicates.memory_search searches by keyword with level, project, status and day filters.memory_project returns a complete project injection section and requires project.memory_find_similar, memory_read and memory_update support duplicate detection, metadata reading and correction.memory_dream consolidates the current window; memory_home views or switches the global data directory.Typing /dream triggers consolidation immediately and bypasses peak-hour suppression. /meow-trust trusts a workspace whose .dsh-meow directory came from an external repository.
The first injection contains long-term soul, user and design-principle records. Keyword hits begin with the second user message and are limited by hitTopK. After compression, the plugin reinjects the long-term snapshot, consulted project data and memories written in the session. Idle dream processing runs in three rounds and is suppressed by default during Beijing time 09:00–12:00 and 14:00–18:00.
DeepSeek Harness compatibility is documented for 0.1.5, including 0.1.5-rc.1. The plugin uses node:sqlite and an esbuild bundle, with no native modules; schemastery and dsh-llm are resolved from the host. Memory is workspace-scoped by default, and externally created .dsh-meow stores are paused until trusted.
Written from the project's own documentation and kept in sync with it. Where the two disagree, the source is authoritative — read the README on GitHub
Same category
vectorize-io
Hindsight, agent memory that learns: long-term project memory with auto recall and retain, knowledge pages, deep reflection, and per-repo memory banks.
No one-line install — this plugin lives inside a larger repository and publishes no npm package.
volcengine
OpenViking memory and context bundle for DeepSeek Harness: pre-step auto-recall and profile injection, session capture, `viking://` URI guarding, and recall/write memory tools backed by an OpenViking server.
Installed cleanly when we ran it
MemTensor
Local-first memory plugin for DeepSeek Harness: layered L1/L2/L3 memory with automatic capture, manual memos_* retrieval tools, and a built-in memory viewer.
No one-line install — this plugin lives inside a larger repository and publishes no npm package.
agentscope-ai
Connects DeepSeek Harness to ReMe's local-first, self-evolving personal knowledge base: automatically captures completed main-agent conversations as user-owned Markdown memory, searches conversations and source material through reme_search with BM25, optional embeddings, and wikilink expansion, and schedules daily memory consolidation.
No one-line install — this plugin lives inside a larger repository and publishes no npm package.
agentscope-ai
Connects DeepSeek Harness to ReMe's local-first, self-evolving personal knowledge base: automatically captures completed main-agent conversations as user-owned Markdown memory, searches conversations and source material through reme_search with BM25, optional embeddings, and wikilink expansion, and schedules daily memory consolidation.
No one-line install — this plugin lives inside a larger repository and publishes no npm package.
zilliztech
Shared Markdown memory for DSH and other coding agents, with automatic capture, pre-step context injection, searchable recall, and memory-to-skill self-evolution through a review panel.
No one-line install — this plugin lives inside a larger repository and publishes no npm package.