Is the agentcrew workflow overkill for solo devs? Looking for honest takes

I built agentcrew around a full BMAD pipeline — BA writes SPEC,

Tech Lead produces DESIGN, PM breaks it into tasks, Dev codes,

QA verifies, QC tests. Six phases, hard gates between them.

For enterprise teams and 5+ person squads, this is a life saver.

No more "half-baked spec → 2 weeks of rework."

But honestly, I'm second-guessing whether it's right for solo devs

or 2–3 person teams. Do you actually need a separate BA persona

when you're the BA, PM, and Dev all at once? Or is the structure

still worth it because AI benefits from clear role boundaries?

Two camps I've heard so far:

🔵 "Even solo, gates prevent me from vibe-coding into a wall"

🔴 "Just let me chat with one AI and ship — this feels like Jira for AI"

Which camp are you in? And if you've tried agentcrew as a solo dev,

where did it start feeling heavy?

Curious to hear real experience, not theory. This will shape the

next release — considering a `--lite` mode that collapses BA+PM

into one, and merges QA+QC.

1 view