Research notes from Nap on secure coding agents, anti-slop engineering, and how AI systems should behave inside real production codebases.
A readable research note on where coding agents can create security risk, and how to keep their work reviewable.
A readable research note on how coding agents remember project intent, use context, and stay useful inside large real-world codebases.
A research note on making coding agents more predictable, reviewable, and useful when they work across real production repositories.
A research note on building review loops where AI-generated code stays readable, bounded, and safe enough for real engineering teams.
28th April, 2026
12th February, 2026
18th May, 2026
27th May, 2026
