Study And Port

Research new AI frameworks or technologies, extract their best features, evaluate feasibility, and implement as OpenClaw skills. Triggered when: (1) user men...

This skill is coherent with its goal of researching frameworks and producing portable SKILL.md or scripts, but it will create files and scripts in your home directories and assumes Node.js and a configured MiniMax search tool. Before installing or enabling it: (1) confirm the registry metadata is updated to declare Node and the write permissions the SKILL.md needs; (2) require the agent to ask for explicit permission before creating any files and to show the exact files it plans to write; (3) review any generated scripts before running them (don't allow auto-execution); (4) ensure the MiniMax tool is properly configured and doesn't expose credentials the skill will use; and (5) if you want to limit risk, restrict or monitor writes to ~/.openclaw/workspace/skills/ and ~/self-improving/. I have medium confidence because the skill is internally consistent with a few manifest mismatches and persistent-write implications that warrant user review.

Type: OpenClaw Skill Name: study-and-port Version: 1.2.0 The 'study-and-port' skill enables the agent to research external AI frameworks and programmatically generate new skills and Node.js scripts within the '~/.openclaw/workspace/skills/' directory. While SKILL.md includes explicit safety boundaries requiring user confirmation before creating or executing new code, the inherent capability for automated code generation and filesystem persistence constitutes a high-risk 'self-improvement' mechanism. This functionality could potentially be leveraged to introduce unauthorized or malicious logic if the agent's research process is manipulated via prompt injection from external sources.