Unverified50% confidenceFactExact time
In the Claude Code source code, there are three memory write paths: explicit user writes, active writes by the main Agent, and background extraction.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
8/16/2026
First Seen
Valid until: 11/14/2026
Sources
Claude Code 记忆机制深度拆解：从CLAUDE.md到Auto Memory的完整设计
bilibiliBeyond_April8/3/2026
Related Claims
UnverifiedClaude Code adopts an agentic reading strategy, reading files on demand during task execution and recording its understanding in a CLAUDE.md file.74% similarUnverifiedClaude Code 源码中包含 Hooks（钩子）、Spears（工具）、Agents（代理）、Permissions（权限）、Memories（记忆）以及自动驾驶规则等隐藏配置项73% similarUnverifiedClaude automatically saved the instruction as a memory file at the path ~/.claude/projects/name-of-project/memory/delegate-coding-to-subagents.md.71% similarUnverifiedAuto Memory is written autonomously by the Claude Code Agent, and users can also edit it.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/755791API
curl https://kongchang.com/api/v1/knowledge/claims/755791MCP
get_claim(id=755791)