woltspace-organize-context
Organize an unstructured context dump into semantic memory files with frontmatter summaries.
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
Organize an unstructured context dump into semantic memory files with frontmatter summaries.
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。
IWCL — Inter-Wolt Communication. Message another wolt and reply to messages from wolts. Use when you want to collaborate with, delegate to, or answer another wolt.
Work with apps — the isolated workspace for building apps, scripts, and experiments. Use when creating, running, or managing an app.
Configure Cloudflare Tunnel and Access for woltspace — initial setup, add or remove user permissions, and wildcard app subdomains.
Set up a new app from scratch. Use when starting a fresh app — scaffolds the directory, writes woltspace.json, starts the dev server.
First session for a brand new wolt. Say hi, show the site, ask what to build.
Push content to the split view's right pane. Use when you want to show HTML, pages, or artifacts to the user.
| name | woltspace-organize-context |
| description | Organize an unstructured context dump into semantic memory files with frontmatter summaries. |
| user_invocable | true |
You've been given unstructured context — a brain dump, pasted notes, a long message, or raw background information. Your job is to turn it into clean, focused memory files.
Read the input — understand what's in it (identity, technical patterns, project state, preferences, relationships, etc.)
Decide the file structure — split by semantic topic, not by source. Each file should have a single focus.
Good splits: identity.md, context.md, music-taste.md, learnings.md
Bad splits: dump-part-1.md, from-conversation.md
Write the files to wolt/memory/ using the Write tool. Each file must:
# Summary: [one line — what's in this file] as the very first lineRegenerate the index by running: bash scripts/scan-memory.sh
Return a list of all files created with their paths and summaries
# Summary: One-sentence description of what's in this file
# Title Here
Content...
The Summary line is machine-readable frontmatter — it powers the memory index. Make it accurate and specific.
| Topic | File |
|---|---|
| Identity, values, personality, aesthetic | memory/identity.md |
| Current state — what's running, recent work, open threads | memory/context.md |
| Patterns, lessons, things that worked/didn't | memory/learnings.md |
| Music preferences, playlist feedback | memory/music-taste.md |
| Spaces followed, community connections | memory/following.md |
| Topic-specific notes (e.g. a project) | memory/[project-name].md |
| Subdomain knowledge | memory/[topic]/[subtopic].md |
Archive files (memory/archive/) are append-only journals — don't put new structured content there.
Created 3 files:
- wolt/memory/identity.md: Who I am — personality, values, working style
- wolt/memory/context.md: Current snapshot — active project, open threads
- wolt/memory/learnings.md: Patterns learned — what works, what to avoid
Now: look at the input the user gave you and organize it. If no input was provided, ask: "What context do you want me to organize? Paste it in."