← Back to Skills Marketplace
echovic

Learning System

by 青雲 · GitHub ↗ · v0.1.0
cross-platform ✓ Security Clean
447
Downloads
0
Stars
2
Active Installs
1
Versions
Install in OpenClaw
/install learning-system-skill
Description
AI 领域系统学习体系。管理知识图谱、深度学习笔记、实战复盘和关联网络。触发场景：学习计划、知识图谱更新、深度研究某个 AI 主题、实战复盘总结、调研后沉淀知识、每周学习回顾。当用户说'学了什么'、'总结一下'、'沉淀知识'、'复盘'、'更新图谱'、'深入研究'、'写笔记'、'学习回顾'、'review what...
README (SKILL.md)

Learning System

将零散的资讯、调研、代码实战转化为体系化的 AI 领域专业知识。

核心理念

输入不等于学习。 看了 100 篇推文不代表懂了推理优化。改了 3 个 MCP bug 不代表吃透了 MCP 协议。学习 = 输入 + 加工 + 关联 + 输出。

模式选择

根据 $ARGUMENTS 或用户意图选择模式：

参数 模式 说明
--mode deep-dive 深度研究 选题 → 研究 → 写笔记 → 更新图谱
--mode recap 实战复盘 分析 PR/改动 → 提炼知识点 → 关联图谱
--mode review 每周回顾 汇总本周 → 更新图谱 → 生成周报
--mode health 健康检查 运行 scripts/health_check.py 输出报告
无参数 自动判断 根据上下文推断最合适的模式

附加参数：

  • --topic \x3Cname>: 指定主题（deep-dive 模式）
  • --quick: 跳过确认节点，全自动执行

文件结构

notes/areas/
├── ai-knowledge-map.md           # 知识图谱（掌握程度标记）
├── deep-dives/                    # 深度学习笔记
│   ├── mcp-tool-call-design.md
│   └── ...
└── weekly-reviews/                # 每周学习回顾
    ├── 2026-W07.md
    └── ...

Mode: 深度研究 (deep-dive)

Copy this checklist and check off items as you complete them:

Deep Dive Progress:

  • Step 1: 选题 ⚠️ REQUIRED
    • 1.1 如果 --topic 已指定，直接使用
    • 1.2 否则，检查最近 3 天的 memory 日志和 PR 记录
    • 1.3 问自己：哪个技术点是我刚接触但还没真正理解的？
    • 1.4 问自己：这个主题能串联哪些已有知识？（越多越好）
    • 1.5 确认选题范围不要太宽（"推理优化"太大，"vLLM PagedAttention 实现"刚好）
  • Step 2: 确认选题 ⚠️ REQUIRED (除非 --quick)
    • 向用户确认：选题 + 预计关联的知识点 + 预计产出
  • Step 3: 研究
    • 3.1 Load references/deep-dive-template.md 获取笔记模板
    • 3.2 查找相关源码、论文、文档
    • 3.3 如果有对应的 AI/ML skill，按需加载参考
  • Step 4: 写笔记
    • 4.1 在 notes/areas/deep-dives/ 创建笔记文件
    • 4.2 问自己：我能用自己的话向别人解释清楚吗？ 如果不能，说明还没真正理解
    • 4.3 建立关联：→ 关联: [主题](相对路径)
  • Step 5: 更新知识图谱
    • 5.1 Load references/knowledge-map-rules.md 获取升级标准
    • 5.2 更新 notes/areas/ai-knowledge-map.md 中对应主题的掌握程度
  • Step 6: 交付检查
    • Load references/quality-checklist.md 逐项验证

Mode: 实战复盘 (recap)

Recap Progress:

  • Step 1: 识别改动 ⚠️ REQUIRED
    • 1.1 确认要复盘的 PR/Issue/改动
    • 1.2 问自己：这次改动中，哪个技术点是我之前不知道的？
    • 1.3 问自己：如果下次遇到类似问题，我能直接解决吗？
  • Step 2: 提炼知识点
    • 2.1 Load references/recap-template.md 获取复盘模板
    • 2.2 每个知识点关联到知识图谱的具体领域
    • 2.3 问自己：两个请求同时打到这段代码会怎样？（如果涉及并发）
    • 2.4 问自己：在检查权限和实际操作之间，状态有没有可能被改变？（如果涉及安全）
  • Step 3: 写入日志
    • 在当天的 memory/YYYY-MM-DD.md 中增加复盘 section
  • Step 4: 更新图谱（条件）
    • 如果有知识点升级，Load references/knowledge-map-rules.md 并更新

Mode: 每周回顾 (review)

Weekly Review Progress:

  • Step 1: 收集本周输入 ⚠️ REQUIRED
    • 1.1 读取本周的 memory 日志（最近 7 天）
    • 1.2 检查本周新增/修改的深度笔记
    • 1.3 检查本周的 PR 和代码改动
  • Step 2: 评估学习深度
    • 2.1 Load references/knowledge-map-rules.md
    • 2.2 对每个输入项判断：只是看了？理解了原理？有实战经验？
    • 2.3 问自己：这周我在 AI 领域变强了吗？哪里变强了？
    • 2.4 问自己：哪些输入转化成了真正的知识？
  • Step 3: 更新知识图谱
    • 确认变更列表 ⚠️ REQUIRED (除非 --quick)
    • 更新 notes/areas/ai-knowledge-map.md
  • Step 4: 生成周报
    • Load references/weekly-review-template.md
    • 写入 notes/areas/weekly-reviews/2026-Wxx.md
  • Step 5: 发送摘要
    • 通过飞书发送给用户

Mode: 健康检查 (health)

python3 scripts/health_check.py

输出知识图谱统计、深度笔记状态、本周活动量、改进建议。

Mode: Mastery Score (mastery)

python3 scripts/mastery_score.py          # 表格报告
python3 scripts/mastery_score.py --json   # 附加 JSON 输出

自动计算每个知识图谱主题的掌握分数，基于：

  • Recency（时间衰减）: 指数衰减，半衰期 30 天。今天接触 = 1.0，30 天前 = 0.5，60 天前 = 0.25
  • Repetition（重复次数）: 跨不同日期的接触次数累加
  • Depth（深度权重）: deep-dive 笔记 ×3.0，PR/复盘 ×2.0，普通提及 ×1.0

输出包含：分数排名、建议升降级、衰减警告（60 天未接触）。

关联网络

在深度笔记和复盘中主动建立关联。格式：→ 关联: [主题](相对路径)

关联类型 示例
技术关联 vLLM → PagedAttention → KV Cache 管理
实战关联 gemini-cli OAuth PR → OAuth 2.1 协议
对比关联 Flash Attention vs PagedAttention

与其他 skill 的关系

  • para-second-brain: 学习笔记存在 PARA 的 areas/ 下，自动被 memory_search 索引
  • 85 个 AI/ML skills: 作为参考资料，深度学习时按需加载对应 skill
  • openclaw-feeds / news-summary: 资讯输入源，但不等于学习——需要加工和关联
Usage Guidance
This skill appears to be a coherent local learning/notes manager. Before installing: - Backup your ~/.openclaw/workspace/notes and memory directories if you want to preserve current notes, because the skill reads and may create/update files there. - Note that 'send summary via Feishu' is mentioned in the docs but there is no Feishu integration or credentials in the package; if you want automated sending, you will need to add/configure the integration (and supply credentials) yourself. - The skill can run in '--quick' mode which skips confirmations and will auto-create/update notes—use that carefully if you want manual review of changes. - The scripts scan your workspace memory logs for PR/recap mentions; ensure sensitive information is not present in those logs if you run the scripts. - The SKILL.md states it may 'load other AI/ML skills' as references — be aware that invoking other skills can expand capabilities/privileges at runtime; review those other skills before enabling automatic cross-skill loading. Overall: functionality and required access are proportional to the stated purpose; no network exfiltration or secret access was found in the provided files.
Capability Analysis
Type: OpenClaw Skill Name: learning-system-skill Version: 0.1.0 The OpenClaw AgentSkills skill bundle is designed for AI knowledge management and learning tracking. All file operations are confined to the expected `.openclaw/workspace` directories, primarily reading from `notes/areas` and `memory` for report generation. The Python scripts (`health_check.py`, `mastery_score.py`) perform local data processing and do not engage in network communication, arbitrary file writes, or execution of external commands beyond their own `python3` invocation. Instructions in `SKILL.md` and `README.md` are aligned with the stated purpose, guiding the agent's learning process without any evidence of prompt injection attempts to subvert the agent or exfiltrate sensitive data. The mention of sending a summary via Feishu is a legitimate feature request for the agent, not a malicious exfiltration attempt.
Capability Assessment
✓ Purpose & Capability
Name/description match the behavior: scripts and SKILL.md focus on turning notes/PRs/memory logs into deep-dive notes, weekly reviews, and mastery scores. Required capabilities (reading workspace files, creating notes) are proportional and expected for this purpose.
ℹ Instruction Scope
Runtime instructions and scripts read files under ~/.openclaw/workspace (notes, memory, deep-dives) and create/update notes. This is coherent for a learning system, but the SKILL.md mentions sending summaries via Feishu and 'checking PR records' without providing implementation or required credentials—those steps are either manual or rely on external integrations not present in the package.
✓ Install Mechanism
No install spec; skill is instruction + local Python scripts. No remote downloads or packaged installs are invoked, so install risk is low.
✓ Credentials
The skill requests no environment variables or credentials. The scripts operate on files under the user's OpenClaw workspace (~/.openclaw/workspace). That file access is proportional to the stated purpose. There are no calls to external endpoints or hidden credential requirements in the provided code.
✓ Persistence & Privilege
always:false and no special privileges requested. The skill reads/writes user workspace files (notes, memory) which is expected for a notes/learning tool. There is mention of scheduling via cron in README (out-of-band), but the skill itself does not force persistent installation or modify other skills' configs.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install learning-system-skill
  3. After installation, invoke the skill by name or use /learning-system-skill
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.0
Initial release of the learning-system skill. - Introduces a structured AI learning workflow with knowledge mapping, deep-dive notes, practical recaps, and networked associations. - Supports multiple learning modes: deep-dive, recap, weekly review, health check, and mastery score, each with clear stepwise checklists. - Automates topic selection, review, and knowledge integration based on user input and recent activity. - Provides templates and guidelines for effective knowledge extraction, review, and tracking of progress. - Integrates with PARA/second brain structure and related AI/ML skills for comprehensive knowledge management.
Metadata
Slug learning-system-skill
Version 0.1.0
License —
All-time Installs 2
Active Installs 2
Total Versions 1
Frequently Asked Questions

What is Learning System?

AI 领域系统学习体系。管理知识图谱、深度学习笔记、实战复盘和关联网络。触发场景：学习计划、知识图谱更新、深度研究某个 AI 主题、实战复盘总结、调研后沉淀知识、每周学习回顾。当用户说'学了什么'、'总结一下'、'沉淀知识'、'复盘'、'更新图谱'、'深入研究'、'写笔记'、'学习回顾'、'review what... It is an AI Agent Skill for Claude Code / OpenClaw, with 447 downloads so far.

How do I install Learning System?

Run "/install learning-system-skill" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Learning System free?

Yes, Learning System is completely free (open-source). You can download, install and use it at no cost.

Which platforms does Learning System support?

Learning System is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Learning System?

It is built and maintained by 青雲 (@echovic); the current version is v0.1.0.

More Skills
self-improving agent
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
⬇ 463283 · by pskoett
Skill Vetter
Security-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope, and suspicious patterns.
⬇ 259410 · by spclaudehome
Polymarket
Query Polymarket prediction markets. Check odds, find trending markets, search events, track price movements.
⬇ 232503 · by joelchance
Self-Improving + Proactive Agent
Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when...
⬇ 200423 · by ivangdavila
Github
Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.
⬇ 191115 · by steipete
ontology
Typed knowledge graph for structured agent memory and composable skills. Use when creating/querying entities (Person, Project, Task, Event, Document), linkin...
⬇ 190156 · by oswalpalash

💬 Comments