Nowledge Community skill
Skills and plugins that connect Claude Code, Cursor, Codex and other agents to Nowledge Mem long-term memory
- 202 ★ on the repository
- Last commit
Extended access
Why this level
- The skills read and write memories in Nowledge Mem, which can hold work decisions and keys from conversations
- Plugins with hooks save sessions and text before context compaction on their own
- The server is expected at `http://localhost:14242`; in a remote setup your archive sits there
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/nowledge-communityDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skills read and write memories in Nowledge Mem, which can hold work decisions and keys from conversations
- Plugins with hooks save sessions and text before context compaction on their own
- The server is expected at `http://localhost:14242`; in a remote setup your archive sits there
This is third-party code. Review the repository files before installing.
What it does
The repository gathers integrations for Nowledge Mem, an app with long-term memory for AI agents. The shortest route is one command, `npx skills add nowledge-co/community/nowledge-mem-npx-skills`, which adds eight skills. `search-memory` looks up past decisions and conversations in normal, deep or progressive mode. `explore-graph` shows a graph of retrieved memories. `read-working-memory` loads the daily briefing at session start. `save-handoff` and `distill-memory` store a session summary and valuable insights, while `check-integration` and `status` verify the connection. Agents with a dedicated integration get separate plugins. The Claude Code plugin adds hooks for startup context, recall during work, automatic session capture and a save before context compaction. There are plugins for Copilot CLI and Droid, a universal plugin for clients that support Agent Plugins, and separate extensions for other environments. A single registry, `integrations.json`, records what each integration does automatically, what its prompts only guide, and what the agent does only when asked.
Who it is for. Nowledge Mem users who work across several agents and want past decisions and session context found without retelling them.
Good fit when
- A bug resembles one you fixed two weeks ago and the agent should find the earlier solution itself
- Every morning the agent needs the project explained again, while a briefing from memory could load at start
- Work moves between Claude Code, Cursor and Codex and each session's outcome must be left for the next agent
- You want to see the links between retrieved memories as a graph
Not a fit when
- Nowledge Mem is not installed or running: the skills only call its server and store no memory themselves
- You want memory without a separate program, such as a rules file in the project: `AGENTS.md` or `CLAUDE.md` is enough
- You need a verbatim import of session transcripts from any agent: the generic package does not promise it and saves summaries
Example request
Search Nowledge memory for how we solved the expiring auth token problem and show the graph of the retrieved memoriesLimitations
- A running Nowledge Mem server at `http://localhost:14242` and the `nmem` CLI are required; the desktop app starts the server on its own.
- The generic `npx skills` package teaches the agent to use memory but guarantees nothing automatic: the model makes the call.
- The `save-thread` skill is kept for compatibility and acts as `save-handoff` in ordinary agents, without saving a full transcript.
- Automatic session capture works where the agent has its own plugin with hooks, such as Claude Code.
How to disable. Delete the `search-memory`, `explore-graph` and other skill folders from the agent's skills directory. The Claude Code plugin is removed with `claude plugin uninstall nowledge-mem@nowledge-community`.
FAQ
How do I install Nowledge Mem skills in Claude Code?
Run `npx skills add nowledge-co/community/nowledge-mem-npx-skills` to get all eight skills. For hooks with automatic session capture, use the native plugin: `claude plugin marketplace add https://github.com/nowledge-co/community`, then `claude plugin install nowledge-mem@nowledge-community`.
Can I install only memory search?
Yes. Add `--skill search-memory` to the install command, and `--skill explore-graph` for the graph. The `--list` flag prints all skills.
How does the Claude Code integration differ from the generic skill package?
The plugin attaches hooks to the agent lifecycle, so context loads and sessions save automatically. The generic package only supplies instructions and the model decides when to search and save.
What is the NMEM_AGENT_ID variable for?
You set it in each child process when an orchestrator launches Codex, Claude Code, OpenCode or other CLIs. Mem uses it to pick the right AI identity. `NMEM_SPACE` overrides the space when needed.
How can I check that the connection to Mem works?
Run `nmem status`: it reports the server connection, its version and the CLI version. The `check-integration` skill does the same from inside the agent and suggests how to set up a native connector.
Discussion
No questions or reviews yet.
Related
SkillsKnowledge-graph memory for agents: Claude Code and Codex plugins, an MCP server and a Python library
A Claude Code skill that pulls memory from Memori Cloud before every non-trivial answer and writes new facts back to it
Shared local memory for AI agents: Claude Code, Codex, Cursor and others find the agreements of past sessions in it and record new ones
Long-term memory for AI agents: a skill and the mindmemos command store facts from conversations and recall them in later sessions
Open infrastructure for agents that learn from experience: session traces, a knowledge base and skill packs via CLI, MCP and API
Local persistent memory for coding agents: hooks record every session, and Claude Code, Codex or Cursor recall past decisions through 54 MCP tools
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Nowledge Community (this card) | yes | 202 | ||
| Mem0 | no | 68,514 | a separate memory layer with its own API and no registry of integrations for a dozen agents | |
| Basic Memory | no | not required | 4,509 | keeps notes as Markdown files on disk, with no memory graph or session hooks |
README in short
The catalog README first describes the `integrations.json` registry and a shared autonomy vocabulary, then lists integrations per tool with install commands. Community Linux packages for Fedora COPR and Arch AUR are listed separately. The skill package README adds a description of each skill and prerequisites.
README badge
Original SKILL.md
--- name: search-memory description: Search Nowledge Mem memories and prior threads for past decisions, procedures, and context with normal, deep, or progressive retrieval; automatically show a focused graph of successful Memory results. --- # Search Memory > AI-powered search across your personal knowledge base using Nowledge Mem. ## When to Use **Strong signals — search when:** - the user references previous work, a prior fix, or an earlier decision - the task resumes a named feature, bug, refactor, incident, or subsystem - the task is a review, regression, release, docs-alignment, or connector-behavior question - a debugging pattern resembles something solved earlier - the user asks for rationale, preferences, procedures, or recurring workflow details - the user uses implicit recall language: "that approach", "like before", "the pattern we used" **Contextual signals — consider searching when:** - complex debugging where prior context would narrow the search space - architecture discussion that may intersect with past decisions - domain-specific conventions the user has established before - the current result is ambiguous and past context would make the answer sharper ## Retrieval Routing Prefer the host's Nowledge Mem MCP tools when exposed: `memory_search` for durable knowledge, `thread_search` for past conversations, and `thread_fetch_messages` for inspecting a matching conversation. Otherwise: 1. Start with `nmem --json m search "query" -n 5` for durable knowledge. 2. Use `nmem --json t search "query" --limit 5` for a prior conversation or exact session history. 3. If a result includes `source_thread`, inspect it once for the specific messages needed with `nmem --json t show <thread_id> --limit 8 --offset 0 --content-limit 1200`. 4. Prefer the smallest retrieval surface that answers the question. For continuation-heavy engineering work, search near the start of the task. Do not wait for the user to literally ask for memory search. Keep the configured endpoint, owner, agent identity, and active space. If the runtime already knows an active lane, pass that scope through the supported MCP parameters or add `--space "<space name>"` to CLI commands. Never infer a space from the current folder or switch owners or spaces to find more results. ## Retrieval Modes - **Normal (default)**: use `memory_search` with `mode="normal"`, or `nmem --json m search "query" --mode normal -n 5`, for concrete facts and bounded recall.