菜单
One single line. Say WHAT the skill does and WHEN to use it, and pack in trigger keywords a user would actually type. Third person. Must stay under 1024 characters and must NOT wrap onto a second line or use YAML folded/literal markers.
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。
基于 SOC 职业分类
Give an AI agent persistent, navigable memory across sessions using a memory-palace structure. Use when building agent memory, when the user says "remember this across sessions", "give the agent long-term memory", "persistent context", "knowledge graph for my agent", or for coding agents, Claude cowork, and ChatGPT projects that must retain state. Reads and writes palace.json as a durable spatial knowledge store the agent reloads each session.
Train mental-imagery vividness and scene construction so memory-palace images are stronger. Use when the user says "my mental pictures are weak/blurry", "I can't visualize", "train my imagination", "help me see images more clearly", or struggles to encode loci. Drills based on imagery research — multisensory detail, manipulation, and scene construction — to make encoding work better.
Turn each target into a vivid, memorable image and place it at a station. Use when the user says "encode these items", "make images for my palace", "help me remember names and faces", "turn this list into pictures", or after memory-palace-architect lays out the rooms. Applies bizarreness, multisensory, exaggeration, motion, and dual-coding; fills the loci[] of palace.json.
Design a memory palace — choose a real or imagined locus, lay a fixed non-crossing route, and space the stations. Use when the user says "design a memory palace", "pick a place to memorize in", "lay out my palace", "map a route for N items", or needs a structure before encoding. Writes the rooms/stations skeleton of palace.json for loci-encoder to fill.
Memorize digits, dates, phone numbers, and playing cards by converting numbers into images. Use when the user says "memorize this number", "remember these dates", "how do I memorize a deck of cards", "PAO system", "Major System", or "Dominic System". Supplies the encoding for numeric targets, then hands the resulting images to loci-encoder for placement.
Teach the method of loci and route the user to the right memory-palace skill. Use when someone says "teach me the memory palace", "how does the method of loci work", "help me memorize X", "I want to remember a list/speech/deck", or is new to spatial memory. The start-here skill that explains the four-level model then dispatches to architect, encoder, number-memory, walk, or spaced-recall.
| name | your-skill-name |
| description | One single line. Say WHAT the skill does and WHEN to use it, and pack in trigger keywords a user would actually type. Third person. Must stay under 1024 characters and must NOT wrap onto a second line or use YAML folded/literal markers. |
One-line statement of the skill's job. Keep the whole file under ~500 lines; push deep material into
references/.
One tight paragraph: the exact situations this skill is for. Be specific — this is what an agent matches against.
Opinionated, ordered guidance. Numbered steps for fragile procedures; prose for open ones.
What this skill must NOT do (the load-bearing guardrails — keep these honest).
references/deep-dive.md — loaded on demand.Built on SIP · mind-palace-agent-skills