promote
Extract key learnings from daily logs and promote them to identity/memory.md. Run automatically via launchd once per day, or manually after a productive day.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Extract key learnings from daily logs and promote them to identity/memory.md. Run automatically via launchd once per day, or manually after a productive day.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
| name | promote |
| description | Extract key learnings from daily logs and promote them to identity/memory.md. Run automatically via launchd once per day, or manually after a productive day. |
Daily log promotion — extract signal from raw conversation logs into long-term memory.
identity/memory.md — always in context, kept under 2500 tokens. Only the essentials.memory/*.md — topic files searched on-demand. Detailed context lives here.
memory/projects.md — project details, architecture, statusmemory/preferences.md — recurring patterns, style choicesmemory/decisions.md — key decisions with reasoningmemory/people.md — people contextmemory/archive-YYYY-MM.md — older entries rotated out of hot memorydaily-logs/ filesidentity/memory.md (hot) and relevant memory/*.md files (cold)identity/memory.md Current State + memory/decisions.mdmemory/decisions.mdmemory/preferences.mdmemory/projects.mdmemory/people.mdidentity/memory.md Active Workidentity/memory.md (hot layer):
## Current State and ## Active Work## Session Logmemory/*.md (cold layer):
<!-- promoted YYYY-MM-DD -->identity/memory.md is over 2500 tokens, archive oldest entries to memory/archive-YYYY-MM.mdRun the morning standup. Use when {{OWNER_NAME}} says "good morning", "standup", or "what's on my plate".
Compress and save the key context from this session to the daily log. Run before ending a long session.
Search through all past daily logs, cold memory files, and conversation history. Use when {{OWNER_NAME}} asks "when did we...", "do you remember...", "what did I say about...", or needs context from a past session.
Bootstrap an UnClaw agent from scratch. Converts a stock Claude Code instance into a persistent personal agent with identity, memory, and scheduling.
Proactive check-in with pre-gathered context. Run automatically via launchd every 30 minutes.
Distill this session's key learnings into long-term memory. Use after completing significant work or when user says "update memory" or "save this session".