loop-benchmarking

Controlled experiments across agentic coding configurations. Same task, one variable, what actually works.
git clone https://git.shiptheloop.com/loop-benchmarking.git
Log | Files | Refs | README

commit 9af3176d11f2f6295abfc41ecc2807dc37c29dd7
parent 5c3c05ad19d7ce1a7fa8554e91ffccceddd0ad94
Author: Brian Graham <brian@buildingbetterteams.de>
Date:   Mon,  6 Apr 2026 06:23:14 +0200

Move artifacts out of Astro public/, fix 13GB node_modules bloat

- Artifacts moved from dashboard/public/artifacts/ to project root
  artifacts/ directory. Astro no longer copies 13GB during build.
- Build time: 60s -> 17s
- Cleaned node_modules from all artifacts (13GB -> 30MB)
- Deploy workflow updated to rsync artifacts separately
- Harness updated to extract to new location

Convex hull scatter plot:
- Replaces individual dots with nested transparent convex hulls per model
- 100% hull (all cells, 8% opacity) + 50% core hull (20% opacity)
- Centroid markers, overlapping regions visible through transparency

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Diff is too large, output suppressed.

Impressum · Datenschutz