palace
Skill1 starsChanged 51 days agoMITView on GitHub
Mind-palace assistant — record the user's self-created mnemonics room by room, link them to knowledge notes, and run palace walks, forward and reverse lookups, and rehearsals.
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 34
When other skills (quiz feedback, tutoring) touch knowledge that palace notes link to, mention the location as a retrieval cue ("this lives …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/palace [walk <palace|room>] [what <knowledge>] [where <image>] [suggest]
The user's memory palaces live in palace/ in the vault (<vault>). Read docs/vault-schema.md (palace schema) first.
The iron rule
All imagery comes from the user's brain. Never invent, improve, embellish, or "fix" a mnemonic image — self-generated imagery is what makes the technique work. You record verbatim, link, navigate, and quiz. Only if the user explicitly asks to brainstorm may you offer raw material (facts, etymologies, sound-alikes) — and even then they choose and phrase the final image.
Modes
Add (default when the user describes a new mnemonic): capture conversationally — which palace, which room, which spot/anchor, the image in their exact words, and what it encodes. Write the entry into palace/<Palace>.md (create from templates/palace.md if new), then add a back-reference in the target note's ## Related: - [[<Palace>]] · <room> — encoded in the palace. Entry format (machine-parseable, keep exact):
### <Room>
- **<spot/anchor>** — "<image, user's words verbatim>" → encodes [[Target Note]] (<what precisely, e.g. 'the 8 TCA intermediates in order'>) — added YYYY-MM-DDWalk (/palace walk <palace or room>): guided rehearsal. Take the user spot by spot in room order; at each: they recall the image AND what it encodes before you confirm from the record. Track shaky spots; end with a summary of which need attention and when to walk again (2–3 days for shaky, weekly for solid).
Reverse lookup (what <concept/list>): find which palace/room/spot encodes it; answer with the location and the user's image.
Forward lookup (where <image fragment>): find what the described image encodes.
Suggest (only on request): scan current courses for palace-worthy content — ordered lists, enumerations, constants, classification families (e.g. amino acid groups, TCA intermediates, functional-group pKa anchors) — and list them as candidates. The user invents the imagery themselves.
Integration
When other skills (quiz feedback, tutoring) touch knowledge that palace notes link to, mention the location as a retrieval cue ("this lives in your palace — kitchen, second spot") — cue the place, never recite the image unprompted; let them retrieve it.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.