Coding agents debug like a junior dev in 2005: they spam print() and loop until the bug disappears or eats your context window.
TL;DR: I gave my agent a debugger and saw it go from 5-6 print-debug cycles down to 1 targeted stop.
I built a small CLI that wraps the Debug Adapter Protocol and exposes basic debugger actions so agents can call them directly. Each command blocks until execution pauses, then returns the full context.
The CLI is just a thin layer. The useful part is a debugging “skill” on top that teaches the agent how to debug methodically with it, define clear hypotheses, validate/invalidate them, etc.
Supports Python, Go, Node/TypeScript, Rust, and C/C++.
TL;DR: I gave my agent a debugger and saw it go from 5-6 print-debug cycles down to 1 targeted stop.
I built a small CLI that wraps the Debug Adapter Protocol and exposes basic debugger actions so agents can call them directly. Each command blocks until execution pauses, then returns the full context.
The CLI is just a thin layer. The useful part is a debugging “skill” on top that teaches the agent how to debug methodically with it, define clear hypotheses, validate/invalidate them, etc.
Supports Python, Go, Node/TypeScript, Rust, and C/C++.