How to Collect Feedback for Coding Agents
Additional_Wish_3619 · reddit · 2026-07-09
While designing a better user feedback mechanism for the local coding agent ATLAS, the author explores how to capture both "overall quality" and "local file-level feedback" in a single run. They argue that scoring an entire run merges good and bad local edits together, leading to trials of an approach combining overall scoring with file-specific overlays.
More from coding & agent
- A Forward Deployed Engineer job really has three stages: audit, evals, deploy — blaizedsouza · 2026-07-22
- 438 sealed tests show coding agents prefer DIY over third-party databases — cramforce · 2026-07-22
- Building a stock research agent turned out to be a prompt-design problem, not a data problem — jwstrategies · 2026-07-22
- Tenable and AWS launch a Black Hat build event for open-source security agents and MCP servers — Dave_Maynor · 2026-07-22
- Discussion: How Much Permission Would You Give Your Background AI Agents? — Total_Drag7439 · 2026-07-22
- Codex helps build Valdiluce, an open-world game with climbing, gliding and gondolas — Dimillian · 2026-07-22