Getting started with AI dev #2 | Hands off project management

This is part two of a three-part series on developing with agents. Part one gave your coding agent a real developer's process, and this part is the plan you hand it to work from. Feed a disciplined engine vibes and you get vibes back, so you grill a fuzzy idea until it's clear, write a spec, slice it into small tickets that name what blocks them, then triage the backlog so the agent always knows what is ready to build. I use Matt Pocock's skills for the whole chain and teach it in two passes, the theory first and then a live build on Balance, my walking-gated RPG, in raw Claude Code. The full write-up is on merc.camp for members. The exact prompt I pasted to install the four skills: Install these 4 skills globally https://github.com/mattpocock/skills/tree/main/skills/productivity/grilling https://github.com/mattpocock/skills/tree/main/skills/engineering/to-spec https://github.com/mattpocock/skills/tree/main/skills/engineering/to-tickets https://github.com/mattpocock/skills/tree/main/skills/engineering/triage ▸ Full write-up on merc.camp: https://merc.camp/experiments/project-management ▸ The whole series (playlist): https://www.youtube.com/playlist?list=PLZaVKZdeMNqo ▸ Part one (start here): https://youtu.be/gxYYrzst1so ▸ Twitch (I stream the dev): https://www.twitch.tv/n3rkmind ▸ Discord: https://discord.gg/zFZE7mfGke ▸ merc.camp: https://merc.camp Chapters 0:00 Intro 1:52 The chain in theory 3:55 Installing the skills 4:07 Building it live on Balance 5:36 Grilling the idea 9:49 Writing the spec 11:03 Slicing into tickets 13:32 Triaging the backlog 14:31 Handing it to the agent 16:41 The results 19:07 The lesson If you're building something and want a team on it, that's what we do at Warden Labs: https://wardenlabs.ca/en