keep-learning-agent
持续学习 Agent - 知识沉淀和经验固化框架。支持学习记录、快速索引、自我修复、经验→模型转化。包含完整模板、索引系统、SOP 流程。让 AI Agent 持续进化,每天进步一点点。
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/chow651/keep-learning-agentWhat This Skill Does
The keep-learning-agent is a sophisticated framework designed to transform your AI agent from a static assistant into a continuously evolving intelligence. It implements a structured knowledge-capture system that ensures every insight, bug fix, or process improvement learned during your sessions is codified, indexed, and made reusable. By utilizing a standardized template system and a dedicated repository structure, the agent facilitates the conversion of ad-hoc experiences into robust, reusable mental models and code snippets. This ensures that no lesson is lost, and your agent improves systematically over time.
Installation
To install this skill, run the following command in your terminal:
clawhub install openclaw/skills/skills/chow651/keep-learning-agent
Ensure your local directory has the required permissions for file creation, as the agent will generate folders like .learnings/ and archive/ to store your institutional memory.
Use Cases
- Engineering Knowledge Management: Document recurring technical debt or unique API patterns so the agent never repeats common mistakes.
- SOP Optimization: Track the efficiency of your internal processes and evolve them through iterative documentation and implementation.
- Self-Repairing Systems: Utilize the included PowerShell scripts to ensure your agent environment remains clean and consistent across sessions.
- Experience Codification: When a repetitive task is mastered, promote it to a reusable module or SOP to automate the workflow for future sessions.
Example Prompts
- "Record this debugging session as a new learning under the 'technical_solution' category; analyze why this error occurred and add a note to the INDEX.md file."
- "Review the pending items in .learnings/LEARNINGS.md and initiate the self-repair sequence for any issues marked as 'resolved'."
- "Summarize all patterns tagged with 'automation' from this week and suggest a new SOP based on these frequent occurrences."
Tips & Limitations
- Consistent Logging: The value of this skill depends on your willingness to categorize learnings correctly using the provided template; poor metadata leads to poor recall.
- Refactor Regularly: Don't just pile up documents; use the
promotedstatus to move stable solutions into thelib/directory. - Security Warning: Since this skill performs file writes to your local drive (including
self-repair.ps1), ensure that you review the scripts it generates before executing them. Treat the agent's generated code as you would any third-party script.
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-chow651-keep-learning-agent": {
"enabled": true,
"auto_update": true
}
}
}Tags
Flags: file-write, file-read, code-execution
Related Skills
context-compressor
Intelligently compress context — conversations, code, logs. Preserve key information while reducing token usage. Auto-detects content type and applies optimal compression.
play-guitar-fretboard
玩转吉他指板 - 快速跳转到吉他指板学习资源网站
faq-distiller
从客服对话、评论、工单或聊天记录中提炼 FAQ,并按用户阶段分类。;use for faq, support, knowledge workflows;do not use for 暴露用户隐私, 替代工单系统.
auto-context
智能上下文卫生检查器。分析当前会话的上下文污染程度 (长对话、主题漂移、噪声累积),建议:continue、/fork、/btw 或新会话。 支持手动触发(/auto-context)和自动触发(响应层实现)。 基于 ArXiv 论文和认知心理学研究的多维度评估体系。
doc-gap-finder
扫描文档目录、标题结构与文件分布,找缺失章节、重复内容和过时区域。;use for docs, audit, knowledge workflows;do not use for 读取无权限目录, 直接修改原文档.