I Built SelfContext So I Could Stop Re-explaining Myself to AI

Описание и применение

SelfContext – локальный слой контекста, позволяющий сохранять и использовать важную информацию в AI‑сессиях без повторного объяснения.

Задача

Every new AI session required me to re-explain my history, goals, preferences, and past decisions, wasting time and losing context continuity.

Результат

SelfContext provides a durable, inspectable, and tool‑agnostic context store that eliminates the need to re‑explain oneself to AI, improving efficiency and continuity across sessions.

Рабочий процесс

• Clone the SelfContext repo locally
• Open the repository root in an AI tool that supports project‑local Agent Skills
• Use the "ingest" command to add new durable items (e.g., resume, decision, fact)
• Use the "query" command to retrieve and reason with existing context
• Use the "checkpoint" command at the end of a discussion to keep only the smallest durable outcome
• Optionally run "upgrade vault latest" after pulling updates

Доказательства

SelfContext is a local context layer for the AI tool I already use.

Источник

Three requests do most of the work: ingest, query, checkpoint.

Источник