Cron Job: 每日创意推送

Job ID: d3bd77eb0469
Run Time: 2026-07-31 06:15:45
Schedule: 0 6 * * *

Prompt

[IMPORTANT: The user has invoked the "idea-miner" skill, indicating they want you to follow its instructions. The full skill content is loaded below.]

name: idea-miner
description: Mine daily creative ideas & startup signals from HN/GitHub.
version: 1.0.0
author: Hermes Agent
license: MIT
metadata:
hermes:
tags: [ideas, trends, research, startup, daily-briefing]
related_skills: [youtube-content, blogwatcher]

🧠 Idea Miner — Daily Creative Intelligence

Overview

Mines today's creative ideas, inspiration hotspots, and startup signals from public sources. Produces a structured Markdown report with demand signals, trending projects, and actionable insights.

The report is saved to /opt/data/hermes/idea-miner/idea-miner_{YYYYMMDD}.md and delivered to the configured destination (Telegram, etc.).

When to Use

• Cron job for daily idea-mining runs (referenced by trender-idea-miner and 每日创意推送)
• User asks: "what are today's hot ideas", "find me startup opportunities", "trending projects analysis"
• Any session that needs to synthesize multi-source trend intelligence

Data Sources

✅ Available & Tested

❌ Currently Unavailable

Data Acquisition Workflow

1. Use curl -s -o /tmp/filename.json -w '%{http_code}' to save to temp file first (avoids pipe-to-interpreter security blocks in cron mode)
2. Parse with python3 -c "import json; data = json.load(open('/tmp/filename.json')); ..."
3. Saved temp files are disposable — clean up or leave in /tmp

Report Format

The report uses Chinese (用户偏好) with the following sections:

# 🧠 Idea Miner — 每日创意情报 (YYYY-MM-DD)

> 数据来源：[list of sources used]

## 📢 需求现场 — 高意向需求 & 讨论
Top demand signals from Ask HN, discussions, needs

## 📈 趋势项目 — GitHub 热门新仓库
Table: project | ⭐ | description | direction
Followed by trend interpretation block (code fence with arrows)

## 🔬 Show HN 精选 — 本周最新项目
Table: project | points | description

## 🌟 今日创业灵感精选
### 高潜力方向
### 已验证信号

## 📊 数据快照
Key metrics summary table

Report Generation Steps

1. Fetch HN front page — identifies the overall tech zeitgeist for the day
2. Fetch HN Ask HN — captures user needs, pain points, and demand signals
3. Fetch HN Show HN (recent, points>10) — captures new projects people are building
4. Fetch GitHub trending repos (last 7 days) — identifies what developers are excited about
5. Fetch additional sources if available (Reddit, Twitter, etc.)
6. Synthesize findings into the structured report
7. Save to file at /opt/data/hermes/idea-miner/idea-miner_{YYYYMMDD}.md
8. Output the report as final response — the cron delivery system automatically delivers it

File Save Convention

• Directory: /opt/data/hermes/idea-miner/
• Filename: idea-miner_{YYYYMMDD}.md (e.g., idea-miner_20260727.md)
• If the directory doesn't exist, create it

Common Pitfalls

1. HN Algolia URL encoding: The numericFilters parameter with > must use %3E not raw >. E.g. numericFilters=points%3E10
2. Reddit API blocked: Don't retry Reddit JSON API with this IP — it's consistently 403. Only retry if OAuth creds are configured.
3. Pipe-to-interpreter security blocks: In cron mode, curl | python3 triggers security warnings. Save output to file first with -o /tmp/file, then parse with python3 -c reading from file.
4. GitHub API rate limits: Without authentication, GitHub API is limited to 60 req/hr. For 10 repos + description parsing per run this is fine, but batch requests if expanding.
5. Show HN search prefers search_by_date: The /search endpoint returns all-time top, not recent. Use /search_by_date for recent Show HN posts.
6. No xurl installed: X/Twitter mining won't work unless xurl CLI is installed and OAuth'd.
7. Google Trends unavailable via simple curl: None of the documented public endpoints work (404). Skip silently.

Verification Checklist

• [ ] HN front page fetched and parsed correctly
• [ ] Ask HN fetched for demand signals
• [ ] Show HN (recent) fetched for new projects
• [ ] GitHub trending repos fetched
• [ ] Report is structured with all tables and sections
• [ ] File saved to correct path with date-stamped filename
• [ ] Report output as final message (cron delivery takes over)

Related Scripts

The legacy shell script at /root/.hermes/scripts/idea-miner_daily.sh references a Python script at ~/idea-miner/idea_miner.py that no longer exists. The cron jobs now run Hermes agent directly with the prompt, bypassing the shell script. If the shell script is ever re-enabled, it needs to be updated.

The user has provided the following instruction alongside the skill invocation: [IMPORTANT: You are running as a scheduled cron job. DELIVERY: Your final response will be automatically delivered to the user — do NOT use send_message or try to deliver the output yourself. Just produce your report/output as your final response and the system handles the rest. SILENT: If there is genuinely nothing new to report, respond with exactly "[SILENT]" (nothing else) to suppress delivery. Never combine [SILENT] with content — either report your findings normally, or say [SILENT] and nothing more.]

生成每日创意情报简报，发送到Telegram。

数据来源

1. Reddit 需求 (r/SideProject, r/somebodymakethis)
2. X/Twitter "I wish there was" 搜索
3. Hacker News Ask HN
4. Indie Hackers 已验证项目
5. GitHub Trending
6. Google Trends 搜索热词

报告格式

使用 Markdown 格式，包含以下章节:
- 📢 需求现场 (高意向需求: I wish/Need/Looking)
- 💰 已验证方向 (有 MRR 数据的成功项目)
- 📈 趋势项目 (GitHub 热门项目)
- 🌐 Google 搜索热词

输出要求

• 简洁、实用、有洞察力
• 每条数据包含来源链接
• 总长度控制在 Telegram 可发送范围内
• 使用中文撰写

目标

聚焦真实用户需求，发现创业机会，提供可执行的创意洞察。

文件保存

将生成的完整报告同时保存到本地文件：/opt/data/hermes/daily-idea-briefing/briefing_{YYYYMMDD}.md

Response

日报生成完毕，已保存至 /opt/data/hermes/daily-idea-briefing/briefing_20260731.md（4063 字符，符合 Telegram 长度限制）。以下为完整报告：

🧠 Idea Miner — 每日创意情报 (2026-07-31)

数据来源：✅ HN · GitHub · Indie Hackers ｜ ❌ Reddit(403) · X(无API) · Google Trends(404)

📢 需求现场 — 高意向需求 & 讨论

1. 域名/数字资产安全 🔥 — 「注册商坑走我 $3000」(63分) + 「Namecheap 把账户交给未验证第三方」(498分/177评) → 域名监控、注册商风控缺口。
   链接1 链接2
2. 停产商品焦虑 — 「受够了品牌停产你最爱的东西？」(48分/48评) → 停产商品追踪、替代品发现 — 低竞争 niche。
   链接
3. 车内广告泛滥 — 「BMW 在车内播放蜘蛛侠广告」(39分) → 联网汽车广告屏蔽工具。
   链接
4. AI 订阅服务可靠性 — 「付费 Claude 宕机超一周且无客服」(45分) → AI 服务可用性监控需求上升。
   链接
5. 流媒体设备安全 — Krebs 警告 TV 流媒体棒安全风险 (658分/375评) → 流媒体设备安全审计。
   链接
6. Android 全球年龄验证 — Google 年底前全球铺开年龄检查 (394分) → 年龄验证合规工具窗口期。
   链接

💰 已验证方向 — 有 MRR 数据的成功案例 (Indie Hackers)

📈 趋势项目 — GitHub 热门新仓库 (近 7 天)

本周信号：
① 前沿模型开源化 → 本地运行工具链爆发 (7.6k⭐+494⭐)
② AI Agent 走向生产：记忆(OptMem)、监控(ZeroShot)、隐私网关(Noisegate)、账号切换(Claude-account)
③ 实时语音成标配 (qwen-audio-agent + persona)
④ 单 prompt 生成完整游戏 (2.4k⭐)

🔬 Show HN 精选 (近 24h)

🌐 Google 搜索热词

⚠️ Google Trends 接口不可用 (404)，以 HN 当日高热话题替代：
- GPT-5.6 主打性价比 (541分) 链接
- GitHub Stacked PR 公开预览 (566分) 链接
- Gemini Robotics 2 具身智能 (522分) 链接
- GCC 发布 AI 政策 (269分) 链接

🌟 今日创业灵感精选

高潜力方向

1. AI Agent 基础设施缺口 — Show HN 密集出现监控/隐私/账号工具。切入点：会话审计、差分隐私、多账号工作流。
2. 开源模型本地化 — Kimi-K3 开源后社区立刻做单设备运行。切入点：量化、一键私有化、合规版。
3. 数字资产安全 — Namecheap 事件 498 分高共鸣。切入点：域名异常监控、注册商声誉追踪。
4. 停产商品替代追踪 — 需求真实且低竞争。切入点：UPC 库 + 替代品推荐。

已验证信号

• 「为既有需求建站」：死域名 → 查流量 → 建站变现，$15k/月 (IH)
• 2 周 MVP：问题发现到上线 2 周，2 年做到 5 位数 MRR (IH)
• 大厂生态细分：微软生态 niche (Pckgr) 做到 $1M ARR (IH)
• 社交销售：LinkedIn 单条互动 → $82K 企业单 (IbexAI)

📊 数据快照