Startup Ideas Bank

A clever plugin shell is not a company until someone cares enough to switch.

AI roast score: 58/100 (D)

The idea

“Everything is a Plugin” sounds elegant until you realize the hard part is not architecture, it’s making developers change tools for a reason stronger than novelty.

The roast

This is a framework looking for a market. You have a technical idea, not a business. The pitch says “open-source agent harness” and “developer preview,” which means you are still in the part where builders admire the machine, not the part where customers pay for it. Your own profile says you sell to developers and enterprises, with subscription and license revenue. That is a hard crowd. They do not buy “everything is a plugin.” They buy reliability, compatibility, and a clear reason to rip out what they already use. The biggest problem is that the core promise is vague. “Everything is a plugin” sounds flexible, but flexibility is not value unless it solves a painful job better than the tools people already have. The pitch admits “THERE WILL BE COMPATIBILITY-BREAKING CHANGES,” which is a giant red flag for enterprise buyers and serious teams. If the product is still breaking itself while asking people to adopt it, your sales cycle becomes a trust problem, not a technology problem. You also have a classic open-source trap: strong developer vibes, weak distribution. “Add the dsh-plugin topic” and “Join DeepSeek Harness Discord” are nice community gestures, but they are not a moat by themselves. Your profile says your moat is switching, speed, and community, which only matters if the harness becomes the default place people build. Right now it is just another repo with a good story and an expensive future.

Red flags

Verdict

Keep building, but do not pretend this is a real business until one painful developer workflow is clearly won and repeatable.

Roast your own startup idea →