DSH Marketplace

engramory

tinqiao-oss/engramory

The Engramory curated-memory discipline as an installable plugin ([npm: dsh-engramory](https://www.npmjs.com/package/dsh-engramory)): a deterministic 200-line / 25 KB cap on the `MEMORY.md` index via `ctx.tools.guard()` — growth denied, a shrinking rewrite always passes — plus the protocol registered as a runtime skill. The store is plain markdown, one file per fact, shared with Claude Code, Codex, Kiro, and OpenClaw.

19512PythonMITSource

Install

Add engramory to DeepSeek Harness

Install into

via GitHub · GitHub source

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.

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 11d agoon DeepSeek Harness 0.2.2-rc.1.

Show it in your README

install verified — dshmarketplace

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

Before you install engramory

  • Source of record: tinqiao-oss/engramory — present in the community registry that DSH's own plugin market installs from.
  • Licensed under MIT.

Security

What installing engramory runs

Build scripts

No build script was blocked in our run.

Source: our sandbox install on DeepSeek Harness 0.2.2-rc.1, 11d ago.

Install hooks

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.

README mentions

shell commands (bash, exec, spawn …); API keys, tokens or passwords

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.

The AI take

What it is — This DeepSeek Harness plugin provides a deterministic capping protocol for the MEMORY.md index.

Who it is for — It is suitable when managing Markdown fact storage in DeepSeek Harness. Users performing writes via shell commands or external editors do not need it, since the index capping does not cover those paths.

Watch out — Installation requires providing an API key or token. Sandbox testing passed in a fresh profile and the plugin was registered. No obvious issues found.

The verdict — I would install it because it denies growth past the index cap for matched edit tools.

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.

What engramory does

Small fact tiles pass through an index gate that blocks growth and allows compacted memory through.

engramory is a DeepSeek Harness plugin that applies a curated, file-based memory protocol to an agent’s local markdown store. It keeps one MEMORY.md index alongside small, human-readable files containing individual facts, with no database, embeddings or server. The protocol is registered as a runtime skill and loaded as standing rules through CLAUDE.md, AGENTS.md or the host’s rules file.

The index is checked through ctx.tools.guard() by a PreToolUse hook for direct Edit, Write and MultiEdit operations. Growth is denied when MEMORY.md would exceed 200 lines or 25 KB; a rewrite that reduces its size is allowed. The discipline also asks the model to deduplicate before writing, update rather than duplicate, remove incorrect facts, and avoid storing information already covered by git, CLAUDE.md or the code. tools/engramory_doctor.py provides a validator.

This suits small-scale, local memory that people can read, edit and diff, particularly where a loaded index is preferable to a database-backed search layer. It is the wrong choice when writes must be globally enforced across shell, MCP, external-editor or sync-client paths: those bypass the hook, and the curation rules remain model-followed. The store assumes one serialised writer. The catalogue flags a terminal surface and credential requirements; the README does not specify which credentials are needed.

engramory documentation

How it behaves

Engramory uses plain markdown as its source of truth: one file per fact plus an always-loaded MEMORY.md index. The live store remains git-ignored. The protocol uses four role-based types: user, feedback, project and reference. feedback is procedural memory and requires Why: and How to apply: lines. Its curation contract covers deduplication before writing, updating instead of duplicating, deleting incorrect facts and excluding information already recorded in git, CLAUDE.md or source code.

The index is intended to stay within both limits below; whichever limit is reached first triggers the cap behaviour.

Limit Behaviour
150 lines / 20 KB Warning threshold
200 lines / 25 KB Compact or ask before writing

Guard and validator

The hard backstop is a PreToolUse hook using ctx.tools.guard(). It checks direct Edit, Write and MultiEdit operations affecting MEMORY.md. A change that would grow the index beyond either cap is denied. A shrinking or compaction rewrite is allowed.

The repository also provides tools/engramory_doctor.py, described as a validator for checking the memory store. The excerpt does not specify its command-line arguments or output format.

Requirements

Engramory is designed for local, file-based memory and does not require a database, embeddings or a server. Its rules are loaded through CLAUDE.md, AGENTS.md or the host’s equivalent rules file. The README describes the project as experimental, version 0.8.0, and assumes a single writer with serialised writes.

Known limits

The hook is not a global write guard. Bash, PowerShell, background Monitor commands, MCP file tools, external editors and sync clients can bypass it. The curation protocol is therefore best-effort and model-followed rather than guaranteed for every task. It is not intended as a mandatory cross-agent memory layer.

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

Alternatives to engramory

coding-agents

vectorize-io

53k

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.

GitHub sourcePython24d ago

AI review

coding-agents

What it is — Hindsight provides long-term project memory management for coding agents.

Who it is for — Developers using AI coding assistants such as Claude Code for code development should install it. If you use a non-Python language for the agent, you can choose other implementations.

Watch out — The plugin requires an API key or token and will execute shell commands. Sandbox testing has not been performed. No obvious pitfalls found.

The verdict — Without sandbox testing results, I would not install it.

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.Read the source
2Details
41k

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

npm packagePython24d ago

AI review

dsh-memory-plugin

What it is — Connects DeepSeek Harness to an OpenViking server for pre-step recall, profile injection, session capture, viking:// URI guarding, and memory read/write tools.

Who it is for — You keep revisiting the same repo in DeepSeek Harness and lose conventions across sessions, with OpenViking as the place those should live. If in-repo docs and rule files already are your context, or you only open throwaway chats, skip it.

Watch out — It installed on a fresh profile and the harness registered it. A reachable OpenViking server is required, and it executes shell commands.

The verdict — I'll turn it on: the product for me is recall happening before the step, not two more tools the agent might ignore.

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.Read the source
1Details
12k

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.

GitHub sourceTypeScript21d ago

AI review

memos-local-plugin

What it is — Hooks DeepSeek Harness to on-device MemOS memory so a turn can be captured, recalled around the next task, queried with memos_* tools, and inspected in a viewer.

Who it is for — You keep driving the same repo inside DeepSeek Harness and each new agent turn starts blind because prior decisions are not stored. If memory has to live in a shared team store rather than on your machine, this plugin is the wrong place for it.

Watch out — There is no published package; you wire it up from a subdirectory of the MemOS repo. You must supply an API key or token, and it executes shell commands.

The verdict — The subdirectory tax only pays off if DSH is already where the work happens—I would not install this as a memory experiment.

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.Read the source
2Source

typescript

agentscope-ai

3.8k

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.

GitHub sourcePythontoday

AI review

typescript

What it is — The plugin integrates DeepSeek Harness with ReMe's knowledge base to automatically capture main-agent conversations as user Markdown memory and schedule consolidation.

Who it is for — When managing main-agent conversations with DeepSeek Harness and wanting to capture completed dialogues as user-owned Markdown memory, this plugin provides integration. For users who do not want to share the workspace with ReMe, independent knowledge base tools offer a different perspective.

Watch out — The plugin requires manually allowing the build script when installing from source. Sandbox testing has not been completed. The plugin executes shell commands.

The verdict — If I am willing to manually allow the build script from source, I would install it because it connects conversations to persistent Markdown memory.

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.Read the source
1Source

ReMe

agentscope-ai

3.8k

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.

GitHub sourcePython12d ago

AI review

ReMe

What it is — Integrates ReMe's local-first self-evolving personal knowledge base with DeepSeek Harness.

Who it is for — When running agents using DeepSeek Harness that require automatic capture of conversations as Markdown memory, this is suitable. If you do not need shared local workspace across agents, you can skip this plugin.

Watch out — No obvious pitfalls found. The plugin will execute shell commands.

The verdict — I would install it, but only if shell command execution is permitted.

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.Read the source
1Source

MemSearch

zilliztech

2.9k

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.

GitHub sourcePython4d ago

AI review

MemSearch

What it is — MemSearch provides shared Markdown memory for DeepSeek Harness coding agents. It supports automatic capture, pre-step context injection, searchable recall, and memory-to-skill self-evolution through a review panel.

Who it is for — When building complex coding agents with DeepSeek Harness that require persistent context across sessions, MemSearch offers value. If your work involves only short-lived conversations without cross-platform memory needs, you can skip it.

Watch out — It will execute shell commands. Sandbox testing has not been done yet; no obvious issues were identified.

The verdict — I would install it because it enables memory to evolve into reusable skills, but only if you are willing to use it via subdirectory installation and accept shell command execution as a trade-off.

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.Read the source
1Source