last30days-openclaw_skill

Overview

This skill researches topics across Reddit, X, YouTube and the web, manages watchlists, generates morning briefings, and queries accumulated research from the last 30 days. It accumulates persistent knowledge in a local SQLite database so insights improve over time. Trigger words route commands into watchlist, briefing, history, or one-shot research modes.

How this skill works

On each invocation the first argument determines the mode: 'watch' for watchlist management, 'briefing' for morning briefs, 'history' to query accumulated knowledge, and any other input for one-shot research. The skill reads mode-specific reference instructions and uses configured APIs (Reddit via OpenAI responses, X via xAI or Bird CLI, YouTube via yt-dlp, optional web search providers) to collect and summarize sources. Results and user preferences are stored under a local SQLite DB and a briefs folder for persistent recall and follow-up queries.

When to use it

• Track emerging conversation or signals on a topic over the last 30 days.
• Create a concise daily morning briefing of recent posts and videos.
• Add or update items on a topic watchlist for ongoing monitoring.
• Run a one-off deep-dive research query across multiple social platforms.
• Query historical research accumulated by the skill for trend context.

Best practices

• Start commands with the mode keyword ('watch', 'briefing', 'history') to ensure correct routing.
• Keep API keys in environment variables or the .env file so source access is prioritized securely.
• Use concise topic phrases for watchlists to avoid overly broad scraping.
• Run the built-in diagnostic command to verify source availability before heavy research.
• Review and edit the stored context file after interactions to capture preferences and adjust source quality notes.

Example use cases

• Add 'quantum computing' to a watchlist and receive a weekly summary of top Reddit and X posts.
• Request 'briefing crypto' each morning to get the latest videos, tweets, and notable web posts.
• Ask 'history climate policy' to surface the last 30 days of saved research and summaries.
• Run a one-shot query 'supply chain semiconductors' to gather immediate cross-platform intelligence.