Create stable agent context
Give long-running AI sessions a shared source of truth for goals, constraints, and decisions.
散らばったプロジェクト情報を、計画、開発、調査、引き継ぎで再利用しやすいメモリーブリーフに整理します。
項目を入力し、生成されたプロンプトをコピーします。
再利用しやすい文脈かを測る実用的な完成度スコアです。
Project memory brief Goal [Add the durable project goal here.] Working context [Add product, audience, constraints, and active workflow context.] Decisions to preserve [Add accepted decisions, rejected options, and reasons.] Risks and watchouts [Add known risks, open questions, and review checkpoints.] Memory rules - Preserve facts, source links, owner names, and dates when they affect future work. - Separate confirmed decisions from assumptions. - When context changes, update this memory instead of appending duplicate notes. - Before acting, check whether a previous decision already answers the question. Response style Track status, blockers, owners, evidence, and next actions.
Give long-running AI sessions a shared source of truth for goals, constraints, and decisions.
Paste the memory prompt at the start of a session so the agent can resume with less setup.
Make assumptions, open questions, and review checkpoints visible before work continues.
It is a reusable brief that tells an AI agent what project facts, decisions, and constraints should persist across sessions.
No. The builder runs in the browser and generates the prompt locally.
Update it after a decision, launch, incident, architecture change, or handoff that future work should remember.