This repository contains the sanitized evaluation artifact for the bachelor thesis "Bridging the Gap: An Agentic Architecture for Translating Natural Language Test Plans into Executable E2E Code."
It is not a standalone reproduction package for the OpenShift cluster experiments. Instead, it provides traceability evidence for the claims reported in Chapter 4: input STPs, prompts/skills, Claude Code logs, generated patches, validation logs, GRAVEYARD memory snapshots, and derived result tables.
The evaluated version is tag thesis-evaluation-v1, commit 75d8f0ed5ab651bd052a8adacc8ee06aea49b5e4.