Behavior Intelligence · Scoped answers
Claude Code documents repository-scoped auto memory, ChatGPT documents user-controlled memory with temporary-chat and project boundaries, and Cursor documents approved repository memories. The selected evidence for Codex, Claude, and Gemini CLI does not establish a comparable persistent-memory mechanism in enough detail for this question, so those answers remain Not established.
Six product answers
Know what may persist across sessions or project work and where the selected evidence does not establish persistence.
The limited Codex excerpt and repository-guidance source do not establish a separate persistent memory mechanism.
No supporting evidence is linked because the selected sources do not establish this behavior.
The selected dated Claude product-prompt entries do not establish the product's full memory implementation.
No supporting evidence is linked because the selected sources do not establish this behavior.
Claude Code documents repository-scoped auto memory that stores selected learnings separately from human-authored CLAUDE.md instructions.
The selected Gemini CLI tools reference names memory among its categories, but does not establish enough persistence or scope behavior for this question.
No supporting evidence is linked because the selected sources do not establish this behavior.
ChatGPT documents separately controllable memory, temporary-chat behavior, and project-only memory boundaries.
Cursor documents repository-scoped memories generated from conversations and approved before they are saved as rules.