tweet-rl-tracker

The tweet-rl-tracker is a specialized agent skill designed for content creators and social media managers who want to treat tweet performance as a quantitative feedback loop. By organizing content creation into a structured reinforcement learning process, this skill allows users to move beyond intuition and identify which hooks, topics, and messaging styles consistently drive engagement. It leverages the Notion MCP for database management and Chrome DevTools for automated visual evidence gathering, providing a systematic approach to iterating on high-performing content.

• Automated documentation of tweet hypotheses, including hook types and target topics.

• Integrated screenshot capture for visual archiving of tweets and video frames using Chrome DevTools MCP.

• Structured data logging of engagement metrics like likes, impressions, and replies to calculate engagement scores.

• Weekly review workflows to compare winners versus losers and extract actionable insights for future content cycles.

• Custom database schema supporting binary "Worked?" indicators for simplified A/B testing and pattern analysis.

• Users should create a dedicated Notion database following the provided schema to log daily tweet performance after 24-48 hours.

• The workflow is optimized for periodic, consistent data entry and weekly retrospective sessions to refine content strategies.

• Inputs require capturing tweet text, link, and metadata; outputs are analyzed via Notion formulas to determine success metrics.

• Be sure to customize the 'Topic' selection field in Notion to align with your specific niche or professional area of interest.

• This system acts as a manual reinforcement learning loop; therefore, the value is derived entirely from the accuracy of your input labels and the discipline of your regular review cycles.