agentsclimarketplace
Sign in

Spaced recall

Skill frankxai/mind-palace-agent-skills/skills/spaced-recall

Schedule spaced repetition over a memory palace so it lasts, and export to Anki. Use when the user says "spaced repetition", "when should I review", "schedule my reviews", "keep this long-term", "export to Anki", "FSRS", or "SM-2". Computes next-due dates from palace-walk grades, writes the recall block of palace.json, and drives durable retention against the forgetting curve.From its SKILL.md

Install
npx -y skills add frankxai/mind-palace-agent-skills --skill spaced-recall

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.5 KB, 578 tokens by cl100k_base, as published. Nobody here has run it

spaced-recall

The durability step. Building encodes; spacing keeps it. Review each locus just before it would fade — not daily, not never.

When to use this

The user wants a palace to survive weeks and months, asks when to review, or wants their loci in a spaced-repetition tool (Anki). Pairs with palace-walk, which supplies the grades.

Instructions

  1. Pick the scheduler:

    • FSRS (Free Spaced Repetition Scheduler) — modern, evidence-led, models per-card stability and difficulty. Default. Targets a desired retention (e.g. 0.9) and spaces accordingly.
    • SM-2 — the classic Anki algorithm; simpler ease-factor model. Use when matching legacy Anki decks or when the user explicitly wants it.

  2. Read the grades. Take per-locus results from palace-walk (got/partial/missed, mapped to the scheduler's grade scale). A missed is a lapse and shortens the next interval sharply.

  3. Compute next-due. Update each locus's recall block: stability, difficulty, reps, lapses, lastReviewed, and the new due (ISO date). New loci start with a short first interval; well-recalled loci stretch out.

  4. Surface today's queue. When asked "what's due", list only loci whose due ≤ today, in route order, and hand them to palace-walk. Never front-load the whole palace — spacing is the point.

  5. Export to Anki when asked — see references/anki-export.md: one note per locus (front = station/cue, back = target + image), preserving FSRS/SM-2 state where possible.

  6. Respect the no-guilt rule. A skipped review is silent; just recompute from the real lastReviewed. No streak penalties, no nagging.

Refusals

  • Don't schedule reviews more often than the algorithm warrants ("review everything daily" wastes effort and isn't spacing).
  • Don't fabricate recall state for loci that were never walked — schedule only what's been tested.

References

  • references/anki-export.md — the Anki export format.
  • spec/MEMORY-PALACE-METHOD.md §4 (forgetting curve, spacing, FSRS/SM-2).

Built on SIP · mind-palace-agent-skills · Memory Palace Method v0.1

What ships with it: 1 file

1.4 KB alongside SKILL.md

references/

Gives 0 of the 12 instructions most study practice skills give in 578 tokens

Counted across 85 of the 97 authors here whose files we hold, read 2026-09-06

  • Ask one question at a timein 17 of 85
  • Give immediate feedback after each answerin 7 of 85
  • Write five to eight quiz questionsin 6 of 85
  • Ask one question per messagein 4 of 85
  • Wait for each answer before continuingin 4 of 85
  • Wait for the user's answer before feedbackin 4 of 85
  • Present multiple-choice via AskUserQuestion with exactly four optionsin 4 of 85
  • Treat pass as merge-readyin 2 of 85
  • Present one word at a timein 2 of 85
  • Verify a story project exists before startingin 2 of 85, across 1 file
  • Clarify the revision pass type firstin 2 of 85, across 1 file
  • Read chapter and continuity context before editingin 2 of 85, across 1 file

Said here and by no other author read

  • Default to the FSRS scheduler
  • Use SM-2 for legacy Anki or explicit requests
  • Read per-locus grades from the palace walk
  • Map grades onto the scheduler's grade scale
  • Shorten the next interval sharply after a missed recall
  • Update each locus's recall block with the new due date

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.