How we build
How we build Sinatra: sandboxing, private-repo security, model choice, and what our own merge data shows.
4 posts
· 5 min read
A coding agent opened 306 PRs on our repos since April. 289 merged.
Nineteen weeks of real data from running a background coding agent on two production codebases: the merge rate, what got closed without merging, and why the human gate is doing most of the work.
How we build · 6 min read
The best model for coding agents depends on the ticket
There is no single best model for coding agents. What differs between frontier models on real tickets, and how to choose per repo without benchmarks.
How we build · 5 min read
How we sandbox AI coding agents
Sandboxing AI coding agents against private code: the threat model, ephemeral sandboxes, audited egress, scoped tokens, and the PR as the boundary.
How we build · 5 min read
Are AI coding agents safe for private repos? 8 questions
Are AI coding agents safe to connect to a private repo? Eight questions to ask any vendor, why each matters, and what a good answer sounds like.
How we build