How do I install the Working Memory skill?

Run npx skillmds add nowledge-co/working-memory in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Working Memory safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: makes network calls. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Working Memory?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Working Memory free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Working Memory?

nowledge-co (@nowledge-co) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Working Memory by nowledge-co — Agent Skill

Name: Working Memory
Rating: 4.5 (4 reviews)
Author: nowledge-co

Working Memory

Load cross-tool Nowledge context at session start. Shows current work, priorities, and unresolved flags. Codex local Memory is separate and does not replace this context; also trigger when resuming or when the user asks what am I working on.

nowledge-co 782e9ca 2.9 KB Updated Sep 9, 2026

File contents

Start with what matters. For full startup context, use Context Bundle: it resolves owner identity, the resolved AI Identity, active scope, active rules, Working Memory, and KFS paths. Working Memory is the lighter daily briefing of active focus areas, priorities, and recent knowledge changes.

Codex local Memory is a separate, local recall layer. Even when it is enabled, use Nowledge Mem for current cross-tool context, exact prior conversations, sourced decisions, and continuity with other agents.

Preferred path

If this session exposes the Nowledge Mem MCP server and you need the full startup contract, prefer read_context_bundle.

Otherwise use read_working_memory for a lightweight daily briefing, or the CLI Context Bundle fallback:

nmem --json context --source-app codex

For only Working Memory:

nmem --json wm read

If the runtime already knows the current project or agent lane, add --space "<space name>". Multi-agent orchestrators can set NMEM_AGENT_ID="<agent-slug>" before launching Codex so the Context Bundle resolves the right AI Identity. Add NMEM_SPACE only when that whole run should override the identity's default space. Use NMEM_HOST_AGENT_ID only for advanced external aliases.

What you'll find

Identity and scope: owner identity, AI Identity, active space, and active rules when using Context Bundle
Focus areas: what you're actively working on, ranked by recent activity
Priorities: items flagged as important or needing attention
Unresolved flags: contradictions, stale information, or items to verify
Recent changes: what shifted in your knowledge base since the last briefing

How to use it

Summarize only the parts relevant to the task. If Context Bundle was loaded, do not separately read Working Memory unless the user asks.
If the task is clearly a continuation, review, release, regression, connector, or prior-decision question, move directly into search-memory after the briefing instead of stopping here.
If exists: false or the command fails, say there is no briefing yet and continue normally.
Share only the parts relevant to what the user is doing now.

When to read

Beginning of a new session
Returning to a project after a break
User asks "what am I working on?" or "what's my context?"

When to skip

Already loaded this session (do not re-read unless the user asks)
User explicitly wants a fresh start
Task is isolated and needs no prior context

Links

---name: working-memorydescription: Load cross-tool Nowledge context at session start. Shows current work, priorities, and unresolved flags. Codex local Memory is separate and does not replace this context; also trigger when resuming or when the user asks what am I working on.--- Start with what matters. For full startup context, use Context Bundle: it resolves owner identity, the resolved AI Identity, active scope, active rules, Working Memory, and KFS paths. Working Memory is the lighter daily briefing of active focus areas, priorities, and recent knowledge changes. Codex local Memory is a separate, local recall layer. Even when it is enabled, use Nowledge Mem for current cross-tool context, exact prior conversations, sourced decisions, and continuity with other agents. ## Preferred path If this session exposes the Nowledge Mem MCP server and you need the full startup contract, prefer `read_context_bundle`. Otherwise use `read_working_memory` for a lightweight daily briefing, or the CLI Context Bundle fallback: ```bash nmem --json context --source-app codex ``` For only Working Memory: ```bash nmem --json wm read ``` If the runtime already knows the current project or agent lane, add `--space "<space name>"`. Multi-agent orchestrators can set `NMEM_AGENT_ID="<agent-slug>"` before launching Codex so the Context Bundle resolves the right AI Identity. Add `NMEM_SPACE` only when that whole run should override the identity's default space. Use `NMEM_HOST_AGENT_ID` only for advanced external aliases. ## What you'll find - **Identity and scope**: owner identity, AI Identity, active space, and active rules when using Context Bundle - **Focus areas**: what you're actively working on, ranked by recent activity - **Priorities**: items flagged as important or needing attention - **Unresolved flags**: contradictions, stale information, or items to verify - **Recent changes**: what shifted in your knowledge base since the last briefing ## How to use it - Summarize only the parts relevant to the task. If Context Bundle was loaded, do not separately read Working Memory unless the user asks. - If the task is clearly a continuation, review, release, regression, connector, or prior-decision question, move directly into `search-memory` after the briefing instead of stopping here. - If `exists: false` or the command fails, say there is no briefing yet and continue normally. - Share only the parts relevant to what the user is doing now. ## When to read - Beginning of a new session - Returning to a project after a break - User asks "what am I working on?" or "what's my context?" ## When to skip - Already loaded this session (do not re-read unless the user asks) - User explicitly wants a fresh start - Task is isolated and needs no prior context ## Links - [Working Memory](https://mem.nowledge.co/docs/advanced-features#working-memory) - [Getting started](https://mem.nowledge.co/docs)

Frequently asked questions