← Back to Skills Marketplace
Memory Pro
by
horizoncove
· GitHub ↗
· v1.0.0
· MIT-0
73
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install yuheng-memory-pro
Description
> 精选层·主动精馏：主动从原始会话和笔记中提取知识，沉淀到长期记忆。。触发词：skill, 优化。
Usage Guidance
This skill is coherent in purpose (saving distilled session notes) but the SKILL.md instructs the agent to create/modify files and run git commits while the metadata declares no required binaries or paths. Before installing: (1) confirm where 'notes/' will be created and that it will be confined to a safe directory or sandbox; (2) ensure you have git installed and understand whether commits will be pushed to a remote (pushes could use remote credentials); (3) ask the skill author to declare required binaries (git) and explicitly state/limit the filesystem paths the skill may write to; (4) consider disabling autonomous invocation or requiring user confirmation for write/commit actions until you trust behavior; and (5) test the skill in a disposable workspace or repo to observe exactly what it writes and commits.
Capability Analysis
Type: OpenClaw Skill
Name: yuheng-memory-pro
Version: 1.0.0
The skill is a knowledge management tool designed to distill session information into structured Markdown notes and version them using Git. It operates within a defined directory structure (notes/) and uses standard commands (git add/commit) to maintain a long-term memory for the agent. No indicators of data exfiltration, malicious execution, or harmful prompt injection were found in SKILL.md or _meta.json.
Capability Assessment
Purpose & Capability
The skill claims to be a memory/refinement layer (writing distilled knowledge into long-term notes). That purpose legitimately requires filesystem write access and a VCS like git, but the skill metadata declares no required binaries or config paths (e.g., no 'git' requirement, no target directory). The absence of these declarations is an incoherence between stated purpose and requested/declared capabilities.
Instruction Scope
SKILL.md explicitly instructs creating/updating notes in notes/<...>.md and running 'git add' and 'git commit'. It also mentions using existing tools (read/write/exec). While writing and committing notes aligns with the skill purpose, the instructions do not constrain target paths, do not limit what may be written, and reference 'exec' which could enable arbitrary command execution if the agent uses it. The instructions also don't state whether commits are local or pushed (remote operations would use remote credentials).
Install Mechanism
This is an instruction-only skill with no install spec or code to download, which minimizes install-time risk.
Credentials
No environment variables or credentials are declared, which is reasonable for simple local note-writing. However, because the skill uses git commands, it will implicitly interact with the user's git configuration and possibly remote credentials (if push is later performed). The metadata should have declared that git and filesystem access are required so users can evaluate credential exposure.
Persistence & Privilege
always:false (good), but the skill performs persistent actions (creating/updating files and committing them). Because the agent can invoke skills autonomously, the combination of automatic triggers and write/exec capabilities could result in unexpected persistent changes to the user's repository or filesystem unless constrained. The skill does not declare path or sandboxing limits.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install yuheng-memory-pro - After installation, invoke the skill by name or use
/yuheng-memory-pro - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of 记忆专家skill (memory-pro):
- Implements an "active distillation" layer for proactively extracting, refining, and saving key knowledge from conversations, notes, and user actions into long-term memory files.
- Clearly defines triggers (end of conversation, discoveries, decisions, corrections, explicit user commands) for when knowledge should be distilled and stored.
- Introduces a systematic workflow for evaluating content value, categorizing, and writing or updating memory files in structured locations.
- Sets standards for high-quality knowledge distillation, including clear summarization, context, applicability, and source tracking.
- Integrates with memory-setup for read operations, coordinates with self-improving-agent's memory architecture, and supports promotion to elite-longterm memory when valuable insights are repeatedly confirmed.
Metadata
Frequently Asked Questions
What is Memory Pro?
> 精选层·主动精馏：主动从原始会话和笔记中提取知识，沉淀到长期记忆。。触发词：skill, 优化。 It is an AI Agent Skill for Claude Code / OpenClaw, with 73 downloads so far.
How do I install Memory Pro?
Run "/install yuheng-memory-pro" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Memory Pro free?
Yes, Memory Pro is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Memory Pro support?
Memory Pro is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Memory Pro?
It is built and maintained by horizoncove (@horizoncove); the current version is v1.0.0.
More Skills