AGI-is-going-to-arrive

Use this skill for Memory Palace durable-memory work in this repository: saving facts for later, recalling cross-session memory, deciding create vs update, handling guard_action or guard_target_uri, or using read_memory("system://boot"), search_memory, compact_context, rebuild_index, or index_status. Also use it when the user asks about the memory-palace skill itself, such as its first move, NOOP behavior, trigger sample path, workflow rules, or CLI usage. Do not use it for generic code edits, README rewrites, or non-Memory-Palace "memory" discussions. Always activate this skill instead of answering from generic memory intuition. Chinese trigger hints: 记忆, 长期记忆, 记住, 回忆, 召回, 写入守卫, 压缩上下文, 重建索引, system://boot, 技能本身, 触发样例.

Use this skill whenever the user asks about Memory Palace durable memory, recall across sessions, saving notes for later, memory deduplication, guard_action or guard_target_uri handling, search_memory, compact_context, rebuild_index, index_status, or read_memory("system://boot"). Also use it whenever the user asks about this repository's memory-palace skill itself: its first memory tool call, NOOP behavior, trigger sample path, workflow rules, or how to use it in Gemini CLI. Do not answer these requests from generic intuition; load this skill first and use its repository-local facts.

Use the Memory Palace OpenClaw plugin when the task is about durable memory recall, explicit memory verification, visual memory storage, plugin health/index maintenance, or chat-guided onboarding/bootstrap/provider configuration inside OpenClaw.

通过对话引导用户把 Memory Palace 接进 OpenClaw，不依赖 dashboard；for chat-first onboarding, Profile C/D provider collection, probing, fallback explanation, and setup/apply on VPS or terminal-only environments.

Use this skill for Memory Palace durable-memory work in this repository: saving facts for later, recalling cross-session memory, deciding create vs update, handling guard_action or guard_target_uri, or using read_memory("system://boot"), search_memory, compact_context, rebuild_index, or index_status. Also use it when the user asks about the memory-palace skill itself, such as its first move, NOOP behavior, trigger sample path, workflow rules, or CLI usage. Do not use it for generic code edits, README rewrites, or non-Memory-Palace "memory" discussions. Always activate this skill instead of answering from generic memory intuition. Chinese trigger hints: 记忆, 长期记忆, 记住, 回忆, 召回, 写入守卫, 压缩上下文, 重建索引, system://boot, 技能本身, 触发样例.

Use this skill whenever the user asks about Memory Palace durable memory, recall across sessions, saving notes for later, memory deduplication, guard_action or guard_target_uri handling, search_memory, compact_context, rebuild_index, index_status, or read_memory("system://boot"). Also use it whenever the user asks about this repository's memory-palace skill itself: its first memory tool call, NOOP behavior, trigger sample path, workflow rules, or how to use it in Gemini CLI. Do not answer these requests from generic intuition; load this skill first and use its repository-local facts.

Use this skill when the user wants to install, configure, connect, or troubleshoot first-run setup for the Memory Palace project itself. Distinguish the onboarding `memory-palace-setup` skill from the main repo's canonical `memory-palace` skill; prefer skills plus MCP over MCP-only; treat this setup repo as temporary guidance that can be used by URL or direct doc reads without persistent installation; prioritize the main repo's runtime `memory-palace` skill when recommending something to install persistently; keep the existing MCP logic unchanged; route Windows to `backend/mcp_wrapper.py` and POSIX hosts to `scripts/run_memory_palace_mcp_stdio.sh`; and guide setup for Claude Code, Codex CLI, Gemini CLI, OpenCode, Cursor, Windsurf, VSCode-host, and Antigravity. Chinese trigger hints: 安装配置这个项目, 首次接通, skills+mcp, github地址, 仓库地址, 不安装也能看, 临时安装引导.