tools-memory
暂无简介
- License 许可
- MIT
- 最近更新
- 2026-11-06
- Stars
- ★ 0
- Forks
- 0
03 SKILL.md 正文
Memory beyond the brain
Chats and past sessions are the aios-brain pack (module 5), not this one. If the member asks to remember or
search past chats across machines, say that first and send them to /aios install aios-brain. Never sync chats with db-sync.
The member's chats and notes are the aios-brain pack (module 5). This pack covers the other two kinds of memory we
use every day: an app's own database that must be the same on every machine, and the clipboard.
The command
Claude Code: "${CLAUDE_PLUGIN_ROOT}/bin/tools-memory". Codex and Cursor: tools-memory (in ~/.local/bin).
| Command | What it does |
|---|---|
doctor |
Turso CLI, db-sync, Paste's MCP server and rich-clipboard: ready, or the install line. |
check <database> |
Read-only. Is the file outside file-sync folders? Does any table number its own rows? Is an app holding it open? |
sample <folder> |
Two practice databases for Acme Studio: one not ready, one ready. |
One database on every machine (db-sync on their own Turso)
The public tools: Turso (turso.tech, free tier, the member's own account) and the public db-sync skill
(git clone https://github.com/skyremote/db-sync ~/.claude/skills/db-sync). The method:
- Check first.
check <their database>. Fix every NOT READY line before anything moves:- move the file out of Dropbox, iCloud, Google Drive, OneDrive and similar folders (file sync corrupts live databases);
- tables that number their own rows need ids made on the machine (ULID or UUIDv7), or two machines will collide. Changing the app's schema is the member's decision; explain it, never do it silently.
- Back it up. Copy the file to a dated
_backupname before setup. - Run db-sync setup on the machine with the real data, following its
references/setup.md: one cloud primary, a local copy per machine, online writes only (never Turso's offline-write mode), a lock around each sync, and credentials in~/.config/db-sync/db-sync.env(chmod 600, never committed, never in a chat). - Prove it: write one test row on machine A,
db-sync now, read it on machine B. Then two writes at once from both machines, and both rows arrive. - Onboard each other machine with
db-sync onboard.
The clipboard (Paste and rich-clipboard)
- Recall: Paste (pasteapp.io, the member's own licence) has a local MCP server. Turn it on in Paste's settings,
then add
http://127.0.0.1:39725/mcp(Claude Code:claude mcp add --transport http paste http://127.0.0.1:39725/mcp). The agent can then search and read clipboard history and pinboards. It stays on their Mac. - Paste formatted: the public rich-clipboard skill puts tables, bold and bullets on the clipboard as rich text
that pastes cleanly into email, Teams and Word
(
git clone https://github.com/skyremote/rich-clipboard-skill ~/.claude/skills/rich-clipboard, MIT).
Rules
- Never sync a database the member has not checked with
check, and never without a backup. - Clipboard history holds passwords and private messages. Search it only for what the member asked for, and never copy what you find into files or chats beyond the answer.
- Free tiers only; ask before anything that costs money.
04 同仓库更多 Skills
your-voice
Help them build THEIR voice skill from their recordings and line edits. Use when they type /your-voice. Never install daniel-voice.
your-ops
Help them componentize THEIR invoice or ops template from a real document they sent. Use when they type /your-ops. Never ship NavAIgate letterhead or KG register facts.
your-brain
Help them start THEIR memory and keep it in step across machines. Use when they type /your-brain. Never install brain-sync or turso-search.
writing-style
Scrub AI slop from prose they are about to ship. Use when they type /writing-style, or when they ask to de-slop a draft. Voice itself is /your-voice — this skill is the editor.
web-scrape
Search the web semantically and pull clean content from pages, including JavaScript-heavy ones. Use whenever you need live web data during a task: finding docs, API references, code examples, articles, papers, or current info (Exa semantic search), or extracting the readable content of a specific URL — especially a JS-rendered, dynamic, or paywalled-by-script page a plain fetch returns empty (Firecrawl). Trigger even when the tools aren't named: "scrape this page", "search the web for", "pull the content from this URL", "find docs/articles about", "extract data from this site", "what does this page say", "research X online", "get me the latest on". Also bundles source clients — keyless Reddit (practitioner sentiment), Academic (OpenAlex papers + Unpaywall free PDFs, citations), Substack (newsletters, long-form), and paid Supadata (YouTube search + multi-platform transcripts) — plus a multi-source deep-research method: reach here for "what do people on Reddit say", "find research papers / citations on", "read this Substack / newsletter", "YouTube transcript", "video transcript", or "do a rigorous deep research pass on X".
transcripts
Pull YouTube, TikTok, Instagram or file transcripts via Supadata. Use when they type /transcripts or paste a video URL and want the words.