Agent Memory System New
v1.0.0OpenClaw Agent 长期记忆系统 - 温度模型 + 自动归档 + 知识提炼。让 AI Agent 拥有持久记忆，自动管理冷热数据，从经验中提炼可复用技能。
⭐ 0· 247·3 current·3 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description match the actual artifacts: scripts create and manage a workspace/memory hierarchy, implement a temperature model, archive >30d files, run nightly reflection and extract skills from lessons. No unrelated credentials, binaries, or external services are required.
Instruction Scope
SKILL.md and the included scripts instruct the agent/user to create directories, run scripts, and add cron jobs. The scripts also automatically modify lesson files (update frontmatter with status and skill-path) when extracting skills — this is consistent with the stated 'skill extraction' feature but is a potentially surprising side‑effect that users should expect.
Install Mechanism
No network downloads or package installs are performed by the included scripts; everything is filesystem operations and cron configuration. No external URLs, shorteners, or archive extraction from remote hosts were observed.
Credentials
The skill does not request environment variables, credentials, or config paths. Scripts optionally respect a WORKSPACE env var but otherwise operate on user home paths (~/.openclaw/workspace). No secrets or unrelated service tokens are required.
Persistence & Privilege
The install routines configure cron jobs and create files under ~/.openclaw/workspace and ~/.openclaw/workspace/skills, giving this skill recurring execution (nightly/weekly). This persistence is coherent with a long‑term memory feature but is an important behavioral privilege you should be aware of.
Assessment
This package is internally consistent with its stated purpose, but it will create/modify files and install cron tasks on your machine. Before installing: 1) read the scripts (memory-gc.sh, nightly-reflection.sh, extract-skill.sh, install.sh) to confirm there is no unwanted behavior; 2) backup ~/.openclaw/workspace and your lessons if you have existing data (extract-skill.sh updates lesson frontmatter and can overwrite); 3) run GC with --dry-run first to see what would be archived; 4) prefer the interactive scripts/install.sh which prompts for confirmation, or manually add the cron entries yourself if you want tighter control; 5) consider running in a test workspace (set WORKSPACE env var) before applying to your real data. If you want, paste the scripts and I can highlight any specific lines to inspect further.Like a lobster shell, security has layers — review code before you run it.
latestvk976851kqsje316h23494g387182bakb
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🧠 Clawdis