Agent QA fails to verify UI fidelity after 3 passes: how do you evaluate design completion?
kevinpiac · reddit · 2026-11-07
A developer describes running a multi-agent team in Firetower (orchestrator, designer, engineer, QA) for a full app redesign. The designer hands off an Artifact with the new design and component prototypes; the engineer implements it; QA verifies.
After 3 QA passes, discrepancies persist. Two root problems: the QA agent can't enumerate all user paths to compare (forgetting important flows), and it can't assert pixel-level fidelity—fonts, colors, and small details drift from the prototype.
The author asks the community how to make the eval loop fully effective and avoid 20 rounds of back-and-forth with agents—a common pain point in agent workflow evaluation.
More from coding & agent
- Agent asked to code an RL experiment turns the context into a rotating goose budget meter — nabla_theta · 2026-11-07
- Veteran C++ dev ships native ComfyUI desktop client, bypassing browser lag — elgeminy · 2026-11-07
- libghostty 1.0 ships, already powering terminals for 12.5M JetBrains users — jedisct1 · 2026-11-07
- Teknium: Context retrieval beats alternatives once sessions exceed 100K tokens — Teknium · 2026-11-07
- Claude Opus 5.5 turns 3-minute task into 3-hour fiasco with 57 self-logged mistakes — PMinervini · 2026-11-07
- Xepak: declarative local builder for REST, RPC, MCP and RAG servers that agents can write DSL for — rumatoest · 2026-11-07