What are the core differences between planning-with-files and OpenSpec?

[JianweiWangs]

Hi， @OthmanAdi Thank you so much for building such an awesome skill like planning-with-files!

I’ve been thinking about the differences between planning-with-files and OpenSpec, and I sometimes find it a bit confusing. Since OpenSpec also focuses on context management, I’m curious about how planning-with-files differentiates itself.

Also, do you think there’s any chance the two could collaborate or integrate with each other in the future?

[OthmanAdi]

Good question, they solve different problems at different stages.

OpenSpec is spec-driven development. You write specs BEFORE code, agree on what to build, then implement. It's great for brownfield projects where you need to define scope clearly before touching existing code. The AGENTS.md file tells the AI how to follow the specs.

planning-with-files is memory during execution. You're already building, and the files track what's happening: what phase you're in, what you discovered, what worked and what didn't. It's the session memory that persists when context fills up.

Think of it this way:

OpenSpec = WHAT to build (planning stage)
planning-with-files = HOW it's going (execution stage)

Can they integrate? Yes. You could use OpenSpec to define your specs upfront, then use planning-with-files to track execution of those specs. The task_plan.md phases could map directly to OpenSpec milestones.

I haven't built an official integration but there's no conflict. They're complementary not competitive.

If you try using both together let me know what patterns emerge. Could be worth documenting.

Ahmad