Stores durable facts in a categorized, plain-markdown vault on disk, alongside your agent's built-in memory.
Memory
Skillhub Cn Daily
Try itDaily Chinese-market skill digest with 8 curated picks and dual-lab security audit.
What it does
Each morning this skill scans SkillHub.cn's 7 rankings, 11 categories, and 6 memory-driven keywords, deduplicating roughly 550 candidates before selecting 8 via a 7-dimension algorithm (trending surge, newcomers, scene match, memory collision, China-first, active developer, optional verification). The final 8 then pass through a 6-dimension AI scoring and dual-lab security audit. Results land as a Markdown brief in three independent destinations — Obsidian inbox, IMA FIM, and Feishu cloud doc — with a 7-day dedup window so already-recommended skills are filtered.
When to use it
- Daily 06:50 cron briefing on Chinese-market SkillHub skills
- Aggregating active SkillHub developers by ownerName
- Receiving a security-audited top-8 list in Feishu or Obsidian
- Filtering out skills already surfaced in the past 7 days
The skill document
SkillHub Daily 国内技能洞察 v7.0
每日扫描 SkillHub.cn 7.5 万+ 技能生态 | 与 ClawHub Daily 互补:聚焦国内适配、活跃开发者、双实验室安全审计
数据源
skillhub CLI 排行榜 + 搜索 API
# 7 种排行榜,每种 100 条
skillhub skill rankings --type hot # 热门榜
skillhub skill rankings --type newest # 最新上线
skillhub skill rankings --type trending # 趋势飙升
skillhub skill rankings --type featured # 编辑精选
skillhub skill rankings --type recommended # 平台推荐
# 11 分类搜索,每种 20 条
skillhub search "知识管理" --json --search-limit 20
skillhub search "办公效率" --json --search-limit 20
skillhub search "内容创作" --json --search-limit 20
# ... 共 11 个分类
# 6 个记忆关键词搜索,每种 20 条
skillhub search "飞书" --json --search-limit 20
skillhub search "Python" --json --search-limit 20
skillhub search "审计" --json --search-limit 20
# ... 共 6 个关键词
# 深度评估(对最终 8 个推荐调用)
skillhub skill evaluation --json # AI 6维质量评估
skillhub skill reports --json # 双实验室安全审计
返回数据结构(每个技能)
{
"slug": "self-improving-agent-cn",
"name": "Self Improving Agent CN",
"ownerName": "zhengxinjipai",
"version": "1.0.0",
"source": "clawhub",
"category": "ai-agent",
"subCategories": [{"key": "agent-memory", "name": "记忆增强"}],
"description": "AI自我改进与记忆系统...",
"description_zh": "AI自我改进与记忆系统...",
"downloads": 37072,
"installs": 9905,
"stars": 49,
"score": 3621.57,
"labels": {"requires_api_key": "false"},
"verified": false,
"claimable": false,
"created_at": 1773023560715,
"updated_at": 1783822732293,
"homepage": "https://api.skillhub.cn/zhengxinjipai/self-improving-agent-cn"
}
分类体系
| category | 中文名 |
|---|---|
| knowledge-management | 知识管理 |
| office-efficiency | 办公效率 |
| content-creation | 内容创作 |
| data-analysis | 数据分析 |
| ai-agent | AI Agent |
| design-media | 设计多媒体 |
| professional | 行业专业 |
| dev-programming | 开发编程 |
| life-service | 生活服务 |
| it-ops-security | IT运维安全 |
| business-ops | 商业运营 |
每日扫描策略
扫描覆盖
| 来源 | 数量 | 用途 |
|---|---|---|
| 7 排行榜 × 100 | 700 | 热门/趋势/精选/推荐 |
| 11 分类 × 20 搜索 | 220 | 长尾覆盖 |
| 6 关键词 × 20 搜索 | 120 | 记忆碰撞精准补充 |
| 合并去重后 | ~550 | 独立技能候选池 |
量化指标
| 指标 | 来源 | 用途 |
|---|---|---|
| downloads | API 字段 | 热度排序 |
| installs | API 字段 | 实际使用量 |
| stars | API 字段 | 用户口碑 |
| score | API 字段 | 平台综合评分 |
| installs/downloads | 计算 | 安装转化率 |
| created_at | API 字段 | 新鲜度 |
| updated_at | API 字段 | 活跃度 |
7 维度推荐算法
D1: trending_surge × 2 — 趋势飙升
同时登上 hot + trending 双榜的技能,score 降序取 Top 2
D2: newcomers × 1 — 新星上线
newest 榜 + created_at 在 30 天内 + installs > 100,installs 降序取 Top 1
D3: scene_match × 2 — 痛点场景匹配
7 大痛点场景库匹配,不同场景各取 Top 1
D4: memory_collision × 1 — 记忆碰撞
3 级权重关键词碰撞:project_memory×3 / topics×2 / user_profile×1
D5: china_first × 1 — 国内优先(最低门槛:50 安装或 10 星)
国内适配信号检测(飞书/微信/钉钉/小红书/抖音等 25 个关键词),候选不足时降级
D6: active_developer × 1 — 活跃开发者
按 ownerName 聚合,活跃度 = 技能数 + 总安装/1000 + 上榜次数×5,推荐最活跃开发者的代表作
D7: tencent_official × 1(可选) — 官方认证
verified=true,候选为 0 时名额分给 D3
7 天去重机制
加载过去 7 天 data/recommended/*.json 中的 slug 集合,推荐时过滤已推荐的技能。
记忆碰撞实现
3 级权重关键词提取:
- project_memory.md (weight=3) — 项目约束、技术决策
- topics.md (weight=2) — 近期任务关键词
- user_profile.md (weight=1) — 用户偏好、技术栈
碰撞算法:对每个技能在 description_zh + name + category + subCategories 中搜索关键词,累加权重。
活跃开发者发现
按 ownerName 聚合所有扫描到的技能,计算活跃度分数,在简报中展示 Top 5 开发者速览(技能数/总安装/上榜次数/代表作)。
使用模式
Interactive 模式
触发词:"SkillHub 日报"、"国内技能推荐"、"有什么新 Skill"
Cron 定时任务
每天北京时间 06:50 执行,TRAE Schedule ID: be17fc27
执行步骤:
1. python scripts/fetch_skillhub_cn.py --output data/snapshots
2. python scripts/daily_recommend.py --data-dir data --skip-eval
3. 三处存放:Obsidian inbox / IMA FIM 知识库 / 飞书云文档
三处存放
| 目的地 | 方式 | 配置 |
|---|---|---|
| Obsidian inbox | 写 Markdown + frontmatter | OBSIDIAN_VAULT_PATH 环境变量(默认 E:\Obsidian\md\inbox) |
| IMA FIM 知识库 | 两步流程(create_note + add_knowledge) | IMA_OPENAPI_CLIENTID / IMA_OPENAPI_APIKEY 环境变量 |
| 飞书云文档 | lark-cli 或 lark-doc skill | 飞书授权 |
三处存放各自独立 try/except,一处失败不阻断其他。
输出物
data/snapshots/YYYY-MM-DD.json— 扫描快照(550+ 独立技能)data/recommended/YYYY-MM-DD.json— 8 个推荐 + 元数据data/recommended/YYYY-MM-DD.md— 中文结构化简报E:\Obsidian\md\inbox\SkillHub-Daily-YYYY-MM-DD.md— Obsidian 本地- IMA FIM 知识库笔记
- 飞书云文档
依赖
- skillhub CLI(
npm i -g skillhub,已全局安装并认证) - Python 3.8+
- 已登录 skillhub auth(
skillhub auth login)
与 ClawHub Daily 互补
| SkillHub Daily | ClawHub Daily | |
|---|---|---|
| 平台 | SkillHub.cn(7.5 万+) | ClawHub.ai(500) |
| 特色 | 🇨🇳 国内优先 / 👤 开发者 / 🔬 安全审计 | 🦞 口碑精品 / 趋势洞察 |
| 评估 | AI 6维评分 + 双实验室审计 | 口碑率 + 活跃度 |
Questions people ask
- How does personalization work?
- It reads three local TRAE memory files — project_memory (weight 3), topics (weight 2), and user_profile (weight 1) — and matches the extracted keywords against skill name, description, category, and subcategories. Only keywords are used; the raw memory content is not transmitted.
- What if one storage destination fails?
- Obsidian, IMA FIM, and Feishu each run inside an independent try/except, so a failure in one does not block the other two from receiving the brief.
- How is the China-first signal detected?
- The scan looks for 25 adaptation keywords such as Feishu, WeChat, DingTalk, Xiaohongshu, and Douyin in skill descriptions, with a minimum threshold of 50 installs or 10 stars. When no candidate qualifies, the slot is reassigned to scene_match.
Related skills
Generate and edit Draw.io, Mermaid, and Excalidraw diagrams from natural language using a structured JSON spec.
Find why your productivity system keeps failing, then apply the smallest fix — capacity math, bottleneck routing, durable local notes.
Join a video meeting as an AI bot with voice, avatar, and screenshare across four operating modes.
Fetch raw ad creative, app, ranking, and revenue data from AdMapix as structured JSON.
Site audit, content writing, and competitor analysis for organic search rankings.
More from edwardwason
Browse all skillsConvert URLs or local files to structured Markdown and save to Obsidian, Feishu Cloud Drive, and Tencent IMA.
Daily curated briefing of 8 notable ClawHub AI Agent Skills, with optional Lark or IMA push.
Convert a Markdown article into 3:4 PNG image cards plus a compressed text script for posting on social platforms.
技能熔炉 — 锻造/评估/改进 Skill。说 技能熔炉 走全流程(含R5改进已有skill);说 技能评估/skill评估/评估技能 只做同类比对+腾讯9维度。可选能力:搜索SkillHub同类技能(通过TRAE内置工具)、修改已有skill文件(仅R5诊断修复路径,需用户确认)。发布环节请用 skill-pu...
Branch a coding session into a new conversation with full context handoff — generate structured handoff doc and startup prompts. Triggers ONLY on '支线任务' / '开...
Generate React design drafts (4-piece set) from content. Invoke for 'design draft'/'设计稿'/'生成页面'/'信息图'/'知识卡片'/'多图配图'. Do NOT use for editing existing code.