LijiAlex/layered-memory

Memory plugin for Claude Code — persistent context from past sessions as markdown. No vector DB.

1Stars on the repository
10Mods indexed here, across every type
2mo agoLast push, which is what freshness is scored on
MITLicence, which decides whether bodies are shown

episode-to-note

01

LijiAlex/layered-memory

Skill Claude CodeCodex

Filing engine — fold a new session feature-episode into ONE existing feature note (rewrite, not append). Used by /memory:build.

1 2mo ago A 33 tokens original MIT

load-memory

02

LijiAlex/layered-memory

Skill Claude CodeCodex

Use when the user asks about something that may be in stored memory, says "load memory", "what do you know about X", "load ", or signals a wrong recall ("wrong memory", "that's not the right context", "reload", "rescan"). Surfaces the relevant stored theme summary as reference.

1 2mo ago A 68 tokens original MIT

summary-to-summary

03

LijiAlex/layered-memory

Skill Claude CodeCodex

Engine B — merge a small CANDIDATE CLUSTER of look-alike feature notes into the fewest coherent notes, preserving all distinct facts. Used by /memory:reconcile (one cluster at a time).

1 2mo ago A 48 tokens original MIT

LijiAlex/layered-memory

Skill Claude CodeCodex

Engine A — classify one Claude Code session and extract a type-aware feature-episode from its (compressed) transcript. Used by /memory:build.

1 2mo ago A 35 tokens original MIT