メニュー
Use Nowledge Mem from WorkBuddy for current context, memory and thread search, durable saves, and WorkBuddy transcript import.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/nowledge-co/community --skill nowledge-memコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
Read the user's Working Memory briefing when current priorities, recent activity, or active focus areas matter.
Read your daily Working Memory briefing to understand current context. Contains active focus areas, priorities, unresolved flags, and recent knowledge changes. In Grok Build, invoke this on the first relevant turn because passive startup-hook output is not model context.
Check Nowledge Mem setup, detect your agent, and guide native connector setup. Use when the user asks about setup, configuration, or when memory tools aren't working as expected.
SOC 職業分類に基づく
SKILL.md を表示中
| name | nowledge-mem |
| description | Use Nowledge Mem from WorkBuddy for current context, memory and thread search, durable saves, and WorkBuddy transcript import. |
Nowledge Mem is the user's current cross-tool memory. Use it to resume with relevant context, recall exact prior work, save durable knowledge, and make WorkBuddy sessions available in other connected AI tools.
At the beginning of meaningful work or when resuming, read Context Bundle through the Nowledge Mem MCP server. The SessionStart hook normally provides it automatically.
CLI fallback:
nmem --json context --source-app workbuddy
If an older CLI rejects context, use:
nmem --json wm read
Do not read both unless the user asks. Use only the parts relevant to the task.
Search Nowledge Mem when the user refers to prior work, resumes a project, investigates a regression, asks why a decision was made, or needs exact conversation history. WorkBuddy-local memory is not a substitute for current cross-tool context or sourced history.
Prefer MCP:
memory_search for durable decisions, preferences, procedures, and learnings.thread_search for prior conversations.thread_fetch_messages only after identifying a relevant thread.CLI fallback:
nmem --json m search "what to recall"
nmem --json t search "conversation to find" --source workbuddy -n 5
Save meaningful decisions, reusable procedures, stable preferences, corrections, and non-obvious learnings. Search first to avoid duplicates.
Prefer MCP:
memory_search for an existing memory.memory_update when the new information evolves it.memory_add for new durable knowledge.CLI fallback:
nmem --json m search "existing concept"
nmem --json m add "content" -t "Title" --unit-type decision -l "label" -s workbuddy -i 0.8
Lifecycle hooks automatically sync the active transcript at PreCompact, Stop, SubagentStop, and SessionEnd. Real transcript capture runs on the WorkBuddy machine; MCP is not a transcript importer.
For a deliberate historical import, preview first:
nmem t sync --from workbuddy --limit 20
Then apply:
nmem t sync --from workbuddy --apply
WorkBuddy transcripts live under $WORKBUDDY_CONFIG_DIR/projects or ~/.workbuddy/projects. The CLI can upload them to either local or remote Mem through the user's configured endpoint.
When setup seems broken:
nmem --json status
If status succeeds but WorkBuddy-specific commands fail, update the same CLI installation before debugging hooks or MCP.
Respect NMEM_AGENT_ID, NMEM_HOST_AGENT_ID, and NMEM_SPACE when present. source_app=workbuddy records provenance; it is not an AI identity.
Use WorkBuddy-owned CODEBUDDY.md and .workbuddy/rules/*.md surfaces for personal behavior. Do not edit installed marketplace files because updates replace them.