agentleFS
erikunhaerikunha/vault-of-loci/.agents/skills/memory-palace/SKILL.md

memory-palace

Skill0 starsChanged 5 days agoMITView on GitHub

Build a memory palace for an ordered list the learner wants to remember, with the learner picking the places, one item at each place, then walk the route and record each place in the learning ledger; use it when the learner wants to remember a sequence, a checklist or ordered steps, and it writes a note, so run it only when asked.

5 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Put one item at each place.

    line 21
  2. Very specific:

    Write the palace note at 50_Review/Palaces/<Name>.md, copied from 90_System/Templates/Memory_Palace.md.

    line 23
  3. Distinctive:

    Read the route back with the script.

    line 30
  4. Very specific:

    After the learner has answered a place, run the script again with --with-items to see the item of that place and check the answer against …

    line 40
  5. Distinctive:

    When the answer is wrong, or the learner asks to try again, use the hint ladder one step at a time: hint, strong hint, partial example, …

    line 47
  6. Distinctive:

    Ask for confidence from 1 to 5 before the reveal.

    line 49

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Tutor mode: the learner attempts first and without notes. The agent never writes the learner's notes or answers. Hints come one step at a time and only after an attempt: hint, strong hint, partial example, answer. Confidence from 1 to 5 is asked before the reveal. The agent never grades an attempt the learner did not confirm.

The palace helps with ordered material and does not replace understanding. The evidence behind it is weaker than the evidence behind retrieval practice.

  1. Find the list the learner means. When they say "this checklist" or name none, look in the track's practice/ and sessions/ notes for an ordered list (a checklist, ordered steps) and name it in one line; ask only when none is there. It must be ordered material: a sequence, a checklist, ordered steps.
  2. Build the route with the learner. In the same reply as step 1, ask for the places. The learner picks the places, because only the learner knows the route. The agent asks for places that are familiar and easy to picture, one place per item, in walking order. The agent does not choose the places.
  3. Put one item at each place. Ask the learner for the items and write them down. An item is one thing, short enough to say in one breath.
  4. Write the palace note at 50_Review/Palaces/<Name>.md, copied from 90_System/Templates/Memory_Palace.md. Fill the table with one row per place, in walking order: Place, Item and Source. The Source column points to the note the item came from. The name of the note gives the ids: the objective is palace-<name> and a place is the item palace-<name>-<place number>. The name is the file name of the note, lowercased, with the underscores turned into hyphens: Home_Route.md gives palace-home-route and palace-home-route-1 for the first place.
  5. Read the route back with the script. It prints the places in order and leaves the item blank, because the item is what the learner has to recall.
python3 -I 90_System/Scripts/palace_test.py --palace 50_Review/Palaces/<Name>.md --places 5
  1. Walk the route with the learner. The agent says one place at a time, in walking order, and the learner says the item for that place, without notes. The agent asks only for the item of the place it just said. It never reads the item out and never shows the table.
  2. After the learner has answered a place, run the script again with --with-items to see the item of that place and check the answer against it.
python3 -I 90_System/Scripts/palace_test.py --palace 50_Review/Palaces/<Name>.md --with-items
  1. When the answer is wrong, or the learner asks to try again, use the hint ladder one step at a time: hint, strong hint, partial example, answer. Never skip a step.
  2. Ask for confidence from 1 to 5 before the reveal. Then reveal the item of that place and grade the learner's words against the note.
  3. Record each place on its own, after the learner confirms the result. Use the objective and the item id of that place. If the learner took a hint, pass --assistance and --hints: an attempt with help stays out of the rate.
python3 -I 90_System/Scripts/record_attempt.py --objective palace-<name> \
  --item palace-<name>-<n> --mode recall \
  [--correct|--incorrect] --confidence <1-5> [--hints <n>] [--assistance <level>] \
  --allow-unknown-objective

The flag --allow-unknown-objective is needed because a palace objective is not in 90_System/Objectives.json. Add the objective to the registry first when the learner wants the palace to show up in the retention report by name.

  1. Show the learner the lines that were written. Never edit or delete a line of the ledger.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.