last30days
研究过去30天的热门话题
在信息噪音中保持对快速变化的AI趋势的敏锐。这个技能会筛选出人们当前真正在点赞和讨论的内容，然后为您的目标工具量身定制可直接复制使用的提示词。
下载技能 ZIP
在 Claude 中上传
前往 设置 → 功能 → 技能 → 上传技能
开启并开始使用
测试它
正在使用“last30days”。 Research Midjourney v6 prompting techniques
预期结果:
Top 5 prompting patterns emerged: (1) multi-shot style references, (2) camera angle keywords, (3) lighting descriptors, (4) aspect ratio tuning, (5) style raw mode. Most upvoted posts emphasize concrete visual anchors over abstract descriptions. Sample prompt structure provided.
正在使用“last30days”。 Best Claude Code skills for git workflows
预期结果:
Most mentioned: /commit (5x), git-worktree skill (3x), /pr creation (3x). The Remotion integration announcement received 16K likes. Key pattern: skills under 500 lines get more adoption. Three specific workflows identified with code examples.
安全审计
低风险Static analyzer detected 484 pattern matches, but 480+ are false positives from markdown documentation (backtick code blocks misidentified as shell execution) and test fixtures. The skill legitimately accesses user-configured API keys (OPENAI_API_KEY, XAI_API_KEY) from ~/.config/last30days/.env and makes HTTP requests to Reddit/X APIs for research. No malicious intent detected. Risk limited to expected behavior for an API-based research tool.
低风险问题 (2)
风险因素
质量评分
你能构建什么
提示词工程研究
通过分析社区分享和点赞的内容，研究特定AI工具的最新提示词技巧
产品推荐发现
按互动信号而非SEO排名，找到人们推荐的具体工具和产品
趋势监控
跟踪AI工具、病毒式内容或新闻等快速变化领域的当前讨论热点
试试这些提示
Research the best [topic] from the last 30 days and summarize what people are recommending
Research [topic] techniques for [target tool] and write copy-paste prompts I can use
Find discussions comparing [product A] vs [product B] from the last 30 days on Reddit and X
Do comprehensive research on [topic] using Reddit, X, and web sources. Include engagement metrics and identify patterns across all three sources
最佳实践
- 同时指定话题和目标工具，以获得最相关的提示词推荐
- 查看互动指标以识别经过社区验证的洞察
- 对于重要话题，使用 --deep 标志进行全面研究
避免
- 不要期望获得超过30天窗口的历史数据
- 不要用于在线讨论较少的话题
- 在根据推荐采取行动前，不要跳过查看原始来源