folded-memory-implementation

SKILLFlusso di lavorocommunity
v0.0.0simbajigegeMITAggiornato 6 g faFonte →

Developer implementation guide for building hierarchical (folded) memory into an Agent. Three-layer architecture where recent turns stay detailed, older content compresses into episodes, and the oldest distills into durable semantic facts. Use when compact-memory-implementation is not retaining enou

Community-submitted skill. Not yet reviewed by the Forge team. Full prompt content may not be available.Request review →
138Stelle del repo
1Client
1Formati
6 g faUltimo aggiornamento
Skill
Autoresimbajigege
Versione0.0.0
LicenzaMIT
CategoriaFlusso di lavoro
Formatiskill.md
PromptNon pubblicato
Compatibilità
Claude✓ Supportato
Cursor
Copilot
ChatGPT
Gemini
Descrizione

Developer implementation guide for building hierarchical (folded) memory into an Agent. Three-layer architecture where recent turns stay detailed, older content compresses into episodes, and the oldest distills into durable semantic facts. Use when compact-memory-implementation is not retaining enough, or when agents need to recall decisions from many sessions ago.

Parole chiave
skillclaude