Browse Comments — Raw (as collected)
Close reading of the corpus at each pipeline stage: raw → clean → relevant → coded.
49
comments matched
· page 3 of 3
Spamming on Linkedin. Is anyone interested in bot posts on Social Media?-----------------------------------------------------------------------------------https://www.linkedin.com/posts/eduardstancu_lately-i-keep-seeing-posts-like-i-built-activity-7445741938579832832-5L7w?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAE3oQMB3ynjQT7BByLkdn4NQVFH_FeRuFI
Enrique Marq this was so well done, and easy to follow, appreciate you sharing!
Impressive level of automation, but the real question is elsewhere: where does human value sit in the loop? If everything becomes generation + distribution, differentiation won’t be production anymore but framing, ideas, and editorial intent. That’s where the human layer becomes critical again.
The "Voice DNA skill" is what makes this genuinely interesting, most social media automation tools produce generic content that sounds like a bot. Teaching the agent to match a specific writing style changes everything.
What's your approach for the CLAUDE.md on this ? With 17 skills, I imagine the system prompt architecture is the real complexity here, how do you handle context window management across that many specialized agents?
Building something similar on the agent instruction side, always curious how others structure their Claude Code setups.
Very nice Enrique Marq I will check the video
Enrique Marq what scheduling tool are you using to move the content from Claude to the social platforms or is the agent in Claude doing this for you?
Coming soon
17 skills and a 3 hour podcast turning into 7 clips without touching it thats wild
Commenting to see if the reply is AI.