AgentEra/Agently-Daily-News-Collector — reverse-engineered prompt
Reverse engineered prompt
Build me a Python command line app that automatically creates a daily news briefing from a topic I type in, like “AI agents” or “climate policy”. It should search the web, choose the most relevant stories, open the pages, summarize them, group them into several columns, then write a polished Markdown report into an outputs folder.
Please use Agently v4 and its TriggerFlow style workflow, with the model settings coming from environment variables or a local env file so I don’t have to hardcode keys. Keep the prompts easy to edit in a prompts folder, and keep the search and browsing parts separate so they can be swapped later.
I want to run it with python app.py, or pass a topic directly from the command line. Add sensible settings for language, search amount, and concurrency. Include logging, a sample settings file, requirements, and clear setup instructions. Look up the current Agently docs online if you need to.
Want more depth? Deep Reverse