一键导入
snipe
// Search Twitter for trending promotional posts about coding/AI agent tools, generate reply drafts with pikiclaw GitHub card, and push results to Feishu doc + bot notification. Does NOT auto-post to Twitter.
// Search Twitter for trending promotional posts about coding/AI agent tools, generate reply drafts with pikiclaw GitHub card, and push results to Feishu doc + bot notification. Does NOT auto-post to Twitter.
This skill should be used to search GitHub for relevant issues, filter them, draft replies to promote pikiclaw using a sub-agent, and publish those replies while tracking already replied issues to avoid duplicates.
Search Reddit (target subs + cross-Reddit search) for recent high-engagement threads about coding agents / Claude Code / AI dev tools, draft peer-builder English reply comments as the pikiclaw builder, push results to Feishu doc + bot. English-first. Does NOT auto-post to Reddit.
This skill should be used when the user asks to start, restart, keep alive, or inspect the local pikiclaw development service, including `npm run dev`, local debug bot startup, dashboard verification, and dev log checks.
This skill should be used when the user asks to "install pikiclaw", "build and install", "deploy locally", "update local binary", "release", or "publish".
| name | snipe |
| description | Search Twitter for trending promotional posts about coding/AI agent tools, generate reply drafts with pikiclaw GitHub card, and push results to Feishu doc + bot notification. Does NOT auto-post to Twitter. |
| user-invocable | true |
| allowed-tools | Bash, Read, Write, Edit, Grep, Glob, Agent, WebFetch, mcp__pikiclaw-browser__browser_navigate, mcp__pikiclaw-browser__browser_take_screenshot, mcp__pikiclaw-browser__browser_snapshot, mcp__pikiclaw-browser__browser_press_key, mcp__pikiclaw-browser__browser_click, mcp__pikiclaw-browser__browser_evaluate, mcp__pikiclaw-browser__browser_type, mcp__pikiclaw-browser__browser_run_code, mcp__pikiclaw-browser__browser_fill_form |
| argument-hint | [keywords] or blank for default |
在 coding agent / AI 开发工具领域的高流量推广帖下回复 pikiclaw,截取流量拿 star。 本 skill 只生成回复草稿推送到飞书,不自动发推(避免封号)。
用户验证过的高效打法:
npx pikiclaw@latest + GitHub 链接历史案例参考:
读取 .pikiclaw/skills/snipe/sniped_posts.txt,避免重复推荐。
使用浏览器工具在 Twitter 搜索近期热门推广帖。
如果用户传入了关键词参数,直接用该关键词搜索。 否则,按以下策略动态搜索。不要死记某个产品名,而是用场景关键词捕获整个领域的推广帖:
搜索关键词(从通用到具体,搜 3-5 组即可):
coding agent tool
AI coding assistant 推荐
claude code 工具
coding agent 开源
remote coding agent
AI agent dashboard
vibe coding 工具
coding agent mobile
AI dev tool launch
搜索操作步骤:
对每个关键词:
https://x.com/search?q={keyword}&src=typed_query&f=top.pikiclaw/skills/snipe/scripts/extract_tweets.js 文件内容browser_evaluate 注入执行该 JS,获取返回的 JSON 字符串text[:80] + url 去重重要:如果某个关键词搜索结果很少或没有推广帖,跳过即可,不要在无效关键词上浪费时间。
从所有提取的推文中,识别 正在推广具体产品/工具 的帖子。
识别推广帖的信号:
必须满足:
sniped_posts.txt 中优先级排序(高到低):
排除:
选出 Top 3-5 条候选帖。对每条候选帖,简要分析它推广的产品与 pikiclaw 的功能交集和差异点。
对每条候选帖,生成回复草稿。
核心原则:读懂原帖在推什么,找到 pikiclaw 相比它最锋利的一个差异点,用最短的文字打穿。
回复格式(极简优先):
{一句差异化,直击原帖产品的短板或 pikiclaw 的独特优势}
npx pikiclaw@latest
GitHub: https://github.com/xiaotonng/pikiclaw
差异化切入角度(从上到下是推荐优先级,挑与原帖产品最对得上的一个):
回复规则:
npx pikiclaw@latest 和 GitHub 链接将候选帖和回复草稿整理为 Markdown 报告:
# Snipe 候选 — {YYYY-MM-DD}
共发现 {n} 条高流量推广帖,以下按推荐优先级排列。
---
## 候选 1: {原帖内容摘要,不超过 20 字}
- **作者**: @{handle}({name})
- **链接**: {tweet_url}
- **数据**: {views} views / {likes} likes / {retweets} RT
- **推广产品**: {product_name} — {一句话描述这个产品做什么}
- **与 pikiclaw 的交集**: {功能重叠点}
- **pikiclaw 的差异优势**: {最锋利的差异点}
### 推荐回复
> {draft}
---
## 候选 2: ...
...
---
## 操作指南
1. 优先回复候选 1(流量最大 / 功能最近),依次往下
2. 在 Twitter 发回复时粘贴 GitHub 链接,Twitter 会自动生成卡片
3. 发完后将推文 URL 追加到 `.pikiclaw/skills/snipe/sniped_posts.txt`
/tmp/snipe_report.mdcd /Users/admin/Desktop/project/pikiclaw && python3 .pikiclaw/skills/snipe/scripts/push_feishu.py --report-file /tmp/snipe_report.md
OK: 开头 → 成功,报告文档 URL 和通知都已发送PARTIAL: → 文档创建成功但通知未发(缺 FEISHU_RECEIVE_ID)ERROR: → 失败,在对话中直接展示报告内容作为兜底将本次所有候选帖 URL 追加到 .pikiclaw/skills/snipe/sniped_posts.txt。
.env 读取(需要 FEISHU_APP_ID、FEISHU_APP_SECRET、FEISHU_CHAT_ID)